SOTES
1386
the n.m.r. spectra of cis- and trans-1,2-diphenylcyclopmpane and lj2-diphenylcyclopeiitane indicate that the 7-values for the aryl hydrogens are larger for the cis isomers than for the trans isomers (less shielding), while
H the
7
I I
values for the methiiie hydrogens (-C-)
are
smaller for the cis isomers. It was also noted that the cis and trans stilbenes did not follow this pattern. The 7-values for the olefinic as well as the aryl hydrogens are both greater for the cis isomer than for the trans isomer. .In a recent publication, Anet2 presented the n.m.r. data for some cis and trans symmetrically disubstituted cyclic derivatives of 2,3-butanediol. These compounds also fit the general pattern shown for the noli-olefinic hydrocarbons. The 7-values of the methyl hydrogens for the cis isomers are larger than for the trans isomers while the 7-values ior the methine hydrogens are smaller for the cis isomers than for the trans isomers. These data are listed in Table I. PROTON
TABLE I CHEMIC4L SHIFTS O F SOME SYMMETRICALLYSUBSTITUTED CYCLICAXD OLEFINICC O X P O C ~ D ~ T
CHa
raryl
rCH
czs-l,2-l>iphenylcyclopropane~ 2 49 6 99 trans-l,2-Diphenylcyclopropanea 2.32 7.32 cis-l,2-Diphenylcyclopentanea 2.62 6 71 trans-l,2-Diphenylcyclopentane~ 2.44 7.11 cis-2,3-Butanediol cyclic carbonate 8.63 4.16 trans-2,3-Butanediol cyclic carbonate 8.59 4.34 cis-2,2,4,5-Tetramethyl dioxolane 8.95 5.83 trans-2,2,4,5-Tetramethyl dioxolane 8 .85 6.62 cis-2-Butene oxideb 9.02 7.30 trans-2-Butene oxideb 8.85 7.43 &-Stilbene oxide“ 2.85 5.83 2.73 6.33 trans-Stilbene oxide“ cis-l,2-Divinylethylene oxided 6.85 trans-l,2-Divinylethyleneoxided 7.15 czs-l,2-Di-(2-furyl) ethylened 3.88 trans-l,2-Di-(2-furyl) ethylened 3.24 2.82 3.51 czs-Stilbene 2.60 3.01 trans-Stilbene 3.72 Diethyl maleated Diethyl fumarated 3.17 3.00 cis-1,2-Dibromoethylene Irans-l,2-Dibromoethylene 3.3s a Values converted from external methylene chloride. All spectra were run as approximately 10% solutions in carbon tetraPrechloride with tetramethylsilane as internal standard. pared by the method of C. E. Wilson and H. J. Lucas, J . Am. Chem. Soc., 58, 2396 (1936). Prepared by the method of H. 0. House, ibid., 77, 3070 (1955). The preparation and chemistry of these compounds will be discussed in a future paper.
We now report the chemical shifts (Table I, entries 9-16) for some symmetrically substituted epoxides and cis- and trans-1 ,2-di-(2-fury1)-ethylene9The chemical shifts for the epoxides correlate with the data for the non-olefinic compounds while the chemical shifts for the furylethylenes correlate with the stilbenes. The n.m.r. spectra of cis- and trans-butene-2 oxide3 (2) F. A. Anet, J . Am Chem. Soc., 84, 747 (1962). (3) It is interesting to note t h a t the methine hydrogens in trans-butene-2 oxide give a quartet while the cts isomer spectrum shows a higher multiplicity indicating t h a t the cw hydrogens are coupled while the trans hydrogens are either not coupled or the coupling IS too small t o observe. This is alao the pattern found for the 1,Z-divinylethylene oxides. The trans hgdrogens give a doublet while the cts hydrogens give a more complicated pattern. I n contrast, the methine hydrogens for both css and trans-2,3-butaned1ol cyclic carbonate are coupled.% The difference is probably due to the existence ot conformers in the case of the trans-2,3-butanediol cyclic carbonate.%
Vol. 67
give 7(CH3)values of 9.02 and 8.85, respectively, while the T(CH) values for the methine hydrogens are 7.30 and 7.43. The 7(CH) values for cis and trans stilbene oxide are 5.83 and 6.33, while T(ary1) values are 2.85 and 2.78, respectively. The .(CH) values for the cis and Irans-ll2-diviny1ethyleneoxide are 6.85 and 7.15, respectively. These values all fit the general pattern shown by the non-olefinic hydrocarbons. The 7-values for the ethylenic hydrogens for cis and trans-lj2-di- (2-furyl)-ethyleiie follow the pat tern set by the stilbenes. The olefinic hydrogens for the cis isomer have a 7-value of 3.88 while the 7-value for the trans isomer is 3.24. It appears that the anisotropes4 of the phenyl and furyl groups are responsible for the olefins not fitting the general pattern shown by the saturated symmetrically substituted compounds. Thus the olefinic hydrogens for the trans isomers are deshielded by the phenyl and furyl groups due to coplanarity.6 A similar situation is found with diethyl maleate and diethyl fumarate6 (entries 21 and 22, Table I), which follow the pattern set by the olefins due to the aiiisotropy of the carbonyl group.5 In contrast the 7-values for cis- and trans-l12-dibromoethylenee are 3.00 and 3.38, respectively. In this instance the bromines in the trans isomer shift the vicinal hydrogen t o higher fields. cis- and transbutene-27 also follow the pattern shown by the noiiolefinic compounds. The spectra of the ethylenic compounds are of particular interest. The publishedl,s spectra of the cis- and trans-stilbenes show the aryl hydrogens of the trans isomer as a multiplet since the aromatic groups are coplanarl~~ while the aryl hydrogens for the cis isomer give a singlet indicating that the rings are not coplanar. Due to the lower order of symmetry of furyl as compared to the phenyl substituent, the suggestion for coliformational preferences is even more definitive in the case of cis- and trans-l,2-di-(2-furyl)-ethylenes. Evidence for the preferred conformations is given by the rather large difference in chemical shift for the Ca hydrogens of the furan ring. The C3hydrogen has a 3.75 7 value for the trans isomer and 3.24 7 for the cis isomer. Thus the Cs hydrogen has different environments for the cis and trans isomers and the molecules must possess different geometries. Acknowledgment.-This work was conducted under Army Contract DA-30-069-ORD-2487, supported by the Advanced Research Projects Agency. (4) L. M. Jackman, “Applications of Nuclear Magnetic Resonance Spectroscopy in Organic Chemistry.” Pergamon Press, London, 1959. ( 5 ) For a discussion of this effect, see ref. 4, Chapter 7. (6) G. V. D. Tiers, “Tables of 7 Values,” Minnesota Mining and b h n u facturing Company, St. Paul, Minnesota, 1958. (7) J. H. Pople, TV. G. Schneider, and H . J. Bernstein, “High-Resolution Nuclear Magnetic Resonance,” McGraw-Hi11 Book Co., New York, P;. Y., Chapter 11, 1959. (8) “High Resolution N M R Spectra Catalogue,’’ Vatian Associates, Palo Alto, California, 1962.
A TRASSPIRATION STUDY OF LITHIUM HYDROXIDE BY JOAN B. BERKOWITZ-MATTUCK AND ALFREDB~;CHLER Aithur D. Little, Inc., Cambridge 40, Massachusetts Recetved J a n u a r y 21, 1968
The increased volatility of lithium oxide in the presence of water vapor was first reported by Brewer
KOTES
June, 1963
and Margrave2 and van Arkel, Spitsbergen, and Heyding.8 The species LiOH(g) was identified mass-spectrometrically by Porter and Schoonmaker,4 who worked a t temperatures between 780 and 900’ and with water vapor pressures of less than 10-2 mm. By analogy with their results on the higher alkali hydroxides, Porter and Schoonmaker suggested that the dimer Liz(OH)z(g) should be an important species a t lower temperatures and higher water vapor pressures. I n a more extensive mass-spectrometric investigation, Berkoomitz, Mexhi, and Chupka5studied the Li,O(c)-H,O(g) equilibrium in the range from 800 to 1100’ at water vapor pressures of 0.04 and 0.16 mm. Under these conditions it was found that dimer accounted for a few per cent of the total lithium hydroxide pressure, and heats of formaitcon of both monomer and dimer were obtained. I n the only other quantitative mvestigation, Smith arid SugdenGaand James and SugdenGb studied the formation of LiOH(g) in flames. The present investigation was undertaken to extend the study of the Li20(c)-H20(g) system by use of the transpiration method to mater vapor pressures beyond the range presently accessible to mass-spectrometric vork. I n this system, two competing equilibria must be considered 1/2H20(g)4- 1/2Li20(c)= LiOH(g) HzO(g)
+ Li,O(c) = LidOH)z(g)
(1)
(1) Supported by the United States Army Bureau of Research under thtx Advanced Research Projects Agency Program, and by the United Stater3 Air Force under a contract monitored by the Missile Deielopment Center, Air Research and Development Command. (2) L. Brewer and J. Margrave, J . Phys. Chem., 59, 421 (1955). (3) A. E. van Arkel, U. Spitsbergen, and R. D. Heyding, C a n . J . Chem., 33, 446 (1955). (4) R. C. Schoonmaker and R. F. Porter, J . Phys. Chem., 6 4 , 457 (1960). (5) J. Berkoaitz. D. J. Meschi, and W. A. Chupka, J . Chem P h u s , 3 3 , 533 (1960). (6) (a) H. Smith and T. h/I Sugden, Proc. Roy Soc. (London), 8 2 1 9 , 204 (1953), (b) C. G. James and ‘r M. Sugden, zhzd., A227, 312 (1955). (7) T. Moeller, “Inorganic Syntheses,” Vol. V, McGran-Hi11 Book Co., New York, N. Y., 1957, p. 1. ( 8 ) ASTM X-Ray Power Data File, Baltimore, Md. (1957). (9) N. W. Gregory and R. Ii. Mohr, J . A m . Chem. Soc., 77, 2142 (1955)
U
F l
0
0
I 20
I
I
40
60
I
80
I
I
100
120
I
140
160
I80
I
FLOW RATE. ML,/MIN.
Fig. 1.-Lithium
oxide weight loss vs. flow rate.
mixture was passed over a weighed sample of LitO(c) contained in a gold boat a t the center of the furnace tube. Carrier gas flow rates were measured with a dibutyl phthalate capillary flowmeter,’O calibrated zn situ with argon by water displacement” from a Mariotte flask12to a precision of &5%. A Richards bottling apparatus was incorporated a t one side of the reaction tube to allow removal of unreacted Li,O( c) for weighing without exposure to air.
Results and Analysis The experimental data obtained are plotted in Fig. 1 in terms of weight loss per unit time w us. flow rate v for a water vapor pressure of 4.68 mm. Weight losses for runs made a t higher water vapor pressures were reduced to the corresponding weight losses a t 4.58 mm. by means of the mass-spectrometric monomer-dimer equilibrium constant6 and are plotted as shaded points. It is clear from the figure that in these experiments, the range of flow rates for which diffusion is significantI3 overlaps the region where the vapor fails to reach saturation. l 4 Therefore, the simple expression generally used to determine vapor pressures from transpiration experiments cannot be applied. If the principal reaction under study can be written as
(2) with equilibrium constants K1 = P L ~ O H / ~ H ~ O and ”~ Kz = ~ L ~ ~ ( o H ) ~ / The ~ H ~dimer-monomer o. ratio in the ~OH gas phase is therefore given by ~ L ~ ~ ( O H ) ~ / P L== ( K 2 / K 1 ) p ~ e ~and ‘ / 2 ,the higher the mater vapor pressure, the greater the importance of the dimeric species. The transpiration measurements to be described were made a t temperatures of 1095 and 1145’K., and water vapor pressures of 4 58 and 19.5 mm. Under these conditions, one estimates from published data5 that Li,O(c) more than 90% of the lithium hydroxide in the equilibrium vapor is Li2(0H)2. Experimental A . Material.-The LizO(c) starting material was prepared by thermal decomposition of lithium peroxide.? The product showed the characteristic X-ray pattern for anhydrous lithium oxide.8 Attempts to synthesize LigO(c) by degradation of Li2C03(c) in a platinum boat a t 1000°2always ended in failure; the product, gray in color, was shown by X-ray fluorescence to contain platinum. B. Apparatus.-A transpiration apparatus of standard de.. signg was constructed of‘ Pyrex, with a reaction tube of fused quartz surrounded by an electrical resistance furnace. Dried, COz-free argon was used au a carrier gas. Water vapor was introduced into the gas stream by bubbling argon through water a t a controlled temperature and then through a series of baffles to avoid mass transfer of liquid droplets. The water vapor-argon
1387
+ HzO(g) Z!Z LL(OHL(g), KZ kf
=
kb
kf ---P L i a ( 0 H ) s _ (3) kb
~
H
~
O
the rate of formation of Li2(OH)2per unit area of sample is given by
where kf and kb are the specific rate constants for the forward and reverse reaction, respectively, and ci is the concentration of species i. If the lithium oxide sample extends from x = 0 to x = I where x measures distance along the reaction tube, and if the concentration of Li2(OH)2(g)can be taken as zero a t x = 0, one obtains atx = 1 CLi2(0H), = C o L i 2 ( 0 H ) e ( l - e - r/v) (5) for a constant partial pressure of water vapor. In ( 5 ) , (10) A. Weissberger, “Technique of Organic Chemistry.” Vol. 111, Part 11, “Laboratory Engineering,” Intersrienoe, Kew York, N. Y., 1957, p. 325 ff. (11) G. W. Smith, I n d . Eng. Chem., Ana2. Ed., 4 , 244 (1932). (12) Scientific Glass Co. Catalog (19591, Bloomfield, N. J., p. 690. (13) U. Merten, J . Phys. Chem., 63, 443 (1959). (14) W. E. Bell and U. Merten, “Kinetic Effects in the Transpiration Method of Vapor Pressure Measurements,” General Atomics Report GA1670, Sept. 9, 1960, San Diego 12, California.
1388
Vol. 67
XOTES
C ' L ~ ~ ( O His ) ~ the
equilibrium concentration of LiZ(OH).?(g) for a given partial pressure of water vapor and y = h$A, where A is the cross-sectional area of the reaction tube. If both diffusion and saturation effects are taken into account and if the lithium hydroxide vapor is assumed to be dimeric, the weight loss measured in the transpiration equipment, the pressure of lithium hydroxide dimer formed, and the flow rate are related by the equation
where T = 298OK. is the temperature a t which the ) ~ equilibrium flow rate was measured, p O ~ i ~ (is0 ~the partial pressure of Liz(OH),(g) for a given partial ( & / A z Z1/A). pressure of water vapor, and p (l/D). I n the latter expression, D is the interdiffusion coefficient for Liz(OH)z(g) and argon, Zz and Az are length and cross-sectional area, respectively, of the isothermal constricted region of the apparatus, and Zl is the length of tubing between the sample boat and the capillary. Equation 6 has three parameters, P ~ L ~ ~ ( O Hp, )~, and y, that may be evaluated from experiment. For v -P 0, w + wo= M L ~ ~ O ~ ~ L ~ ~ (and O Hfor ) J Rv T+ , a, w --t w, = A~L~,OPOL~,(OH)~Y/RTP. Over the range of flow rates studied a point (w*, v*) must exist for which (1 = (1 - e-pv*), and hence w* = ( W O W , ) I / ~ , while w*/v* = !~I!ZL~~(OH)~/RT. In order to draw the curves g. hr.--l and shown in Fig. 1, wm,1095 = 2.6 X wm,l145 = 4.0 X 10-4 g. hr.-l (where the second subscript refers to the temperature in OK.) were estimated from the experimental data, since it is clear that in practice the asymptotic value of w is reached a t finite flow rates. From the graph the values of 2 X g. hr.-l and 1.4 X g. hr.-l were chosen for w0,1096 and w0,1146 respectively. Values of w* were computed and trial values of v* selected until the curves shown in Fig. 1 were obtained. For both of these curves v* = 1.5 ml./min. This procedure is expected to give ~ ~ ) z to within a factor of two. values of p O ~ i z ( accurate With the parameters given, equilibrium constants K2,1~95 = 0.105 and KZ.1145 = 0.357 were obtained for reaction 2. Equation 6 permits an estimate of the efficiency with which gaseous water molecules react with LizO(c) to form Lie(OH)z(g). Unsaturation effects will become important a t flow rates v such that
+
kiAl
v=y=K
(7)
For the present experiment one has approximately and K = K2,1100 = 1.3 x 1 = 3 cm., A = 5 so that (7) becomes
v
=
1.15 X 104hml. set.-'
(8)
If the rate of formation of Li2(OH)?(g)is written as the product of the number of collisions 2 of water molecules per LizO unit and a collision efficiency a, then, for a surface density p = 10l6sites em.-* one obtains
kf
=
aZ - = 2 x 108 P
(9)
Since unsaturation effects are apparent for flow rates
greater than 5 ml. min.-l, the collision efficiency a must be less than Discussion In the present experiments, the reaction rate becomes a limiting factor a t moderate flow rates, and observed weight losses in the plateau region of Fig. 1 will depend on the surface area of the sample. This accounts for the poor reproducibility between experiments in that region. On the other hand, a t low flow rates, weight losses are inconveniently small, and the flow rates are difficult to monitor and maintain. Therefore, the measured equilibrium constants are probably accurate to no better than a factor of two. Even so, they are an order of magnitude higher than those reported by Berkowitz, Meschi, and Chupka. In the present v mole/ work the flow of water vapor is 0.29 X min., where v is the volumetric flow rate in cc. min.-l. The water vapor pressures and effusion orifice dimensions used in the mass-spectrometric experiments corand 25 X low6mole respond to flow rates of 6 X min.-1. It is therefore just possible that the difference between the two sets of measurements is real, and that solid-gas equilibrium was not established in the earlier measurements. It was not practicable to extend the transpiration measurements over a sufficiently wide temperature range to determine meaningful second-law enthalpies. KOweight loss was observed in a 10 hour experiment at 1045OK. with a sTater vapor pressure of 4.58 mm. and a carrier gas flow rate of 4.0 ml. rnin.-l. At 1170°K., on the other hand, extensive devitrification of the quartz reaction tube rendered the data suspect. It is possible, however, to analyze the results at 1095 and 114.j°K. by the third-law method,16 to obtain an enthalpy of formation of Lin(OH)e(g). Thermal functions for LizO(c) and HzO(g) have been reported in the literature.16 For Liz(OH)zfree energy functions and absolute entropies were calculated using the vibrational frequencies and bond lengths estimated by Berkowitz, lleschi, and Chupka.6 From the data a t 1095 and 1145OK., respectively, one calculates the heat of formation of Li2(OH)z at OOK. as -183.6 and -186.1 kcal./mole. The heat of formation calculated from the paper of Berkomitz, Meschi, and Chupka is -180 kcal./mole, to be compared to an average value from this work of - 185 kcal./mole. (15) G. N. Lewis and M. Randall, "Thermodynamlcs," Second Edition, revised b y L. Brewer and K. Pimer, hIcGraw-H111, Ken York, N . Y . , 1961, p. 178. (16) JANAF Thermochemical Tables, Quarterly Supplement No. 4 , The Thermal Laboratory, Dow Chemical Co., Dee. 31, 1961.
T H E EXCESS VOLUME OF BISARY MIXTURES OF SORMAL ALKANES B Y JOS6 D.
G6YEZ-IBAfiEZ AND
CHIA-TSGNLTU
Hall Laboratory of Chemistry, Weslsgan Uniberszty, Middletoun, Connecttcut Received J a n u a r y 6 , 1968
The so-called "principle of congruence" mas formulated by Brqinsted and Koefoedl in order to represent the excess free energies of binary mixtures of n-alkanes. At a given temperature and pressure, the excess thermody(1) J. N. Br$nsted and J. Koefoed, Kgl. Danske Videnakob. Selakab., Mat.-fus. Medd., 22, No. 17,1 (1946).