Mass Spectrometric Study of the Sublimation of Lithium Oxid - The

Chem. , 1959, 63 (5), pp 644–648. DOI: 10.1021/j150575a002. Publication Date: May 1959. ACS Legacy Archive. Cite this:J. Phys. Chem. 63, 5, 644-648...
1 downloads 0 Views 594KB Size
644

J. BERKOWITZ, W. A. CHUPKA,G. D. BLUEAND J. L. MARGRAVE

Vol. 63

MASS SPECTROMETRIC STUDY OF THE SUBLIMATION OF LITHIUM OXIDE1 BY JOSEPH BERKOWITZ, WILLIAMA. CHUPKA AND GARYD. BLUEAND JOHN L. MARGRAVE Argonne National Laborator Lemont, Illinois Department of Chemistry, University of #isconsin, Madison, Wisconsin Received June 10, 1068

Lithium oxide sublimes mainly by decomposition to the elements at 1400"K., but an appreciable partial pressure of LizO(g) in equilibrium with the solid is indicated by mass spectrometric studies of Liz0 sublimation from a platinum Knudsen cell. From absolute pressures and estimated thermodynamic functions, the following heats were obtained: LizO(s) = LizO(g), AHoo = 104 f 5 kcal./mole; LiO(g) = Li(g) O(g),. DoO L 83 kcal./mole (3.6 e.v.). LiO(g) is of very minor importance as a vapor species, but LiONa(g), formed when Na is present as an impurity, appears to have considerable stability.

+

Introduction The nature of the gaseous species subliming from solid alkali metal oxides has been the subject of considerable speculation and experimentation. Brewer12from a consideration of the data available in 1951, decided that there was no evidence for the existence of important alkali metal oxide gaseous molecules except possibly Liz0 and LiO. Theoretical calculations by Brewer and Masticka treating the M2O molecule as a linear assemblage of three ions, suggest the CszO, Rb20, K20 and NazO should vaporize by decomposition to the elements with no appreciable contribution due to MzO gaseous molecules. These calculations do indicate the possible importance of LizO, but the uncertainty is large. Experimental determinations of the vapor pressures of Li20 and Na20 have been made by Brewer and Margrave4 using the Knudsen effusion technique. The observed pressures of both substances agreed within experimental error with the pressures expected for simple decomposition to the metal gases and Oz. However, when the same investigators measured the volatility of Lip0 in streams of argon and oxygen, no decrease in volatility upon changing from argon to oxygen at one atmosphere was observed, thus demonstrating that the partial pressure of LinO gaseous molecules must be comparable to that of the elements in the saturated vapor of LizO. I n this work on lithium oxide, a mass spectrometer has been used to investigate the vapor in thermodynamic equilibrium with solid LizO inside a Knudsen cell. Observation of ion masses and measurement of ion currents then provide a means of identifying the species in the vapor and obtaining thermodynamic data for the gaseous equilibria inside the Knudsen cell. With this experimental arrangement it is possible to obtain more unambiguous information than that which may be obtained from either Knudsen weight loss or flow type experiments. Experimental The mass spectrometer and associated Knudsen cell assembly have been described in detail earlier by Chupka and (1) Work performed under the auspices of the U. S. Atomic Energy Commission. (2) L. Brewer, C l e m . Reus., Sa, 1 (1953). (3) L. Brewer and D. F. Mastick, J . A m . Chem. Soc., 73, 2045 (1951). (4) L. Brewer and J . Margrave, J . Chem. Phus., 59, 4 2 1 (1955).

Inghram,b and Porter, Schissel and Inghram.6 In this experiment, the Knudsen cell was made of platinum and was contained within an Inconel oven. It was necessary to contain the sample completely in platinum since the Inconel acts as a reducing agent as was evidenced in the earliest runs without a platinum lid where the vapors were in contact with the Inconel cover. The temperature of the cell was measured with a calibrated optical pyrometer, which was sighted through a Pyrex window and a slot in the radiation shields onto a black-body hole in the cell. A correction was made for the transmission of the window. The vapor effusing from the Knudsen cell entered directly into the ionization chamber of the mass spectrometer where the neutral atoms and molecules were bombarded with electrons of controlled energy and thus ionized. An electron multiplier was used for ion detection. The output of this multiplier was measured by a vibrating reed electrometer and recorded by a strip chart recording potentiometer. This provided ready measurement of ion currents as low as amp. The lithium oxide was prepared by heating lithium carbonate to about 1000" for 1.5 to 2 hours and pumping o f f the Cog. Brewer and Margrave4 discuss the chemical and physical properties of the product obtained in this manner. A small amount of sodium impurity (about 0.3%) threatened to thwart efforts to determine the existence of a gaseous LiO bqth occur a t mass 23. molecule since zsNa+ and %i160+. To overcome this difficulty, some lithium oxide used in one run was prepared from 6Li metal as 26Li 2H20 = 26LiOH HZ (NH4)&03 26LiOH = 6LiZCOs 2"s 2H20 6LizCOa = SLizO COZ The 6Li160+ ion has mass 22, a relatively "clean" position coni ared t o mass 23 with its continuously varying amount of 23&a+ due to depletion of the sodium impurity with time.

+

+

+

+ +

+

.

Experimental Results

A. Identification of the Gaseous Species.When the decomposition reaction of the lithium carbonate apparently was complete and the pressure due to the evolved COz gas was reduced, a mass scan was made and the species produced by electron bombardment of the effusing vapor were identified. Peaks a t masses 7, 30 and 32 were shown to be 'Li+, 'Li2160+and 160z+, respectively. A subsequent investigation of the isotopic structure of the peaks due to the two lithium isotopes provided a positive check on the identity of the lithium-containing species. Using normal lithium oxide, a peak at mass 23 was found to be due to Na+, as mentioned before. However, when one run was made with a sample of lithium oxide en(5) W. A. Chupka and M. G. Inghram, THIEJOURNAL, S9, 100 (1955). (6) Porter, Schissel and M. G. Inghram, J . Chem. Phys.. 28, 339 (1955).

b

May, 1959

*

MASSSPECTROMETRIC STUDYOF

riched with the 6Li isotope, a peak at mass 22, attributable to "il60+ was observed. Appearance potential curves for Li+, LiO+ and Li20+ are shown in Fig. 1. With the known ionization potential (5.36 e.v.) of gaseous lithium atoms as a standard, it was determined that A.P. (LiO+) = 9.0 f 0.2 e.v. and A.P. (LizO+) = 6.8 0.2 e.v. A peak observed at mass 46 was a t first attributed to LizOa+,but later identified as LiONa+ produced by ionization of LiONa which was formed by the reaction of the Na with the LizO. Evidence for the identity of the species observed a t mass 46 was the fact that the intensity at mass 46 followed the decline in intensity of the sodium impurity as the activity of the latter diminished. In addition, the mass 46 peak was much more intense during the initial runs under reducing conditions which is contrary to the effect one would expect for an Liz02 species. B. The Equilibrium: LizO(s) = 2Li(g) l/zOz(g).-It was possible to calculate the equilibrium constant for the decomposition reaction and hence the equilibrium partial pressures of gaseous lithium and oxygen over the temperature range investigated by making use of thermodynamic data which were available in the literature. For such calculations, the heat of formation and entropy for formation of solid Li20 a t 298°K. have been given by Johnston and Bauer.' The relative enthalpy, HTO - HzssO and the relative entropy, STO S298O, of the solid were evaluated by extrapolating the equations of Shomate and Cohens from 1045"K., the highest temperature a t which their data were obtained, to 1600°K. All the necessary functions for lithium and oxygen were obtained from the tabulation of t,hermodynamic properties of the elements by Stull and Sinke.g The results of these calculations are presented in Table I. The partial pressure of a particular species is related to the observed ion intensity by the equation P = kI+T where P = vapor pressure, k = proportionality constant, I + = positive ion current, and T = absolute temperature.'O It is possible to evaluate the proportionality constant in the ion currentpressure equation (and therefore the absolute partial pressure of IL gaseous species inside the Knudsen cell) from a sensitivity calibration with a substance of known pressure and an estimate of the relative ionization cross In this experiment it was assumed that the effusing vapor had the composition of the solid. The relstive ionization cross sections u then were evaluated for Li and 0 2 by comparing the observed ion intensities of these species on the detector and using the factor ml/m2'/2 as a correction for speed of transit across the ion chamber. The observed ion intensity of LizO+ was converted to pressure of LizO by assuming that a(LizO) = 2u(Li) i/2a(02)and

THE

SUBLIMATION OF LITHIUM OXIDE

645

+' rr

+

-

+

(7) H. L. Johnston and T. W. Bauer, J . A m . Chem. Soc.. 73, 1119 (1951). (8) C. H. Shomste and A. J. Cohen, ibid., 71, 285 (1955). (9) D. R.Stull and G . C. Sinke, "Thermodynamic Properties of the Elements." Advances in Chemistry Series No. 18 (American Chemical Society, 1956). (10) W. A. Chupks and M. G. Inghram, J . Chem. Phye., 21, 371 (1953). (11) M. G. Inghrsm, W. A. Chupka and Porter, ibid., as, 2161 (1955).

5 Fig. 1.-Ionization

10 15 Electron energy. efficiency curves for lithium oxide vapor species.

comparing to IO*+and the previously calculated pressure of 0 2 . In a like manner the pressure of LiO was obtained from the ion intensity of LiOf by . assuming that a(Li0) = u(Li) 1 / 2 ~ ( 0 z ) Taking ~ ( 0= ~ 1, it) was found that v(Li) = 1.1, a(LizO) = 2.7, and a(Li0) = 1.6. The calculated cross sections of Otvos and Stevenson12 show that u(Li)/ ~(OZ) = 1.26. Insofar as these calculations may be assumed to approximate the true cross sections, experimental evidence is thus obtained for the absence of reduction and for the validity of the above assumption that the vapor and solid have very nearly the same composition. C. The Equilibrium: LizO(s) = LizO(g).-Two independent methods were employed to extract the heat of sublimation of the LizO from the experimentally observed ion currents of Li20+. By making use of the ion current-pressure relationship, P = kI+T, and the Clausius-Clapeyron equation

+

a value of AHo,,b which is independent of the proportionality constant, k , may be found from the slope of the curve obtained when plotting In (I+T) us. 1/T. A Clausius-Clapeyron plot for the sublimation of Liz0 is shown in Fig. 2. The slope of this plot yielded a heat of sublimation of 108.3 kcal./mole-' over the temperature range 12481534"K., and a heat of sublimation of 115.9 kcal./ mole-' at absolute zero. The alternative procedure was to use k , the calibration constant evaluated as previously discussed, to determine the absolute pressure of LizO. The (12) J. W . Otvos and D. P. Stevenson, J . Am. Chem. Soc., ?E, 546 (1956).

J. BERKOWITZ, W. A. CHUPKA, G. D. BLUEAND J. L. MARGRAVE

646

Vol. 63

TABLFI I T (OK.)

1000 1050 1100 1200 1300 1400 1500 1600

I/T

x

LizO(s) = 2LiW 104

10.000 9.524 9.091 8.333 7.692 7.143 6. 667 6.250

+ 1/202(g)

log Kea

log Po2

-30.32 -28.07 -26.02 -22.45 -19.44 -16.88 -14.67 -12.74

-12.61 -11.71 -10.89 9.46 - 8.26 - 7.23 - 6.35 - 5.58

-

PLl(r.trn)

P02btrn)

2.46 x 1.95 X 1.29 X 3.47 x 5.50 x 5.89 X 4.47 x 2.63 X

9.84 x 7.80 x 5.16 x 1.39 x 2.20 x 2.36 x 1.79 X 1.05 X

10-13

lo-'* lo-"

10-10 10-9 10-7

10-1s lO-l2

10-11 10-9 10-8 10-7

.

estimated stretching bond constant of 6 X lo6 dynes cm. -l yielded a ground-state vibration frequency of 1450 crn.-I. The LizO molecule was assumed to be bent with an Li-0-Li bond angIe of 105" and a bending force constant of 0.65 X lo6 dynes cm.-', obtained by comparison to similar molecules such as DzO. The valence-bond approxim a t i ~ n was ' ~ used to determine the three frequencies of vibration. The frequencies 1270, 740 and 1650 cm.-l were used in the computations. Table 111 presents the computed thermodynamic functions for the LiO and Liz0 gaseous molecules which were used in the absolute pressure treatment of the mass spectrometric data.

11.0

10.0

F cr

5 0.0

i .

M

*

3

TABLE I1 Li,O(s) = LinO(g) 8.0

7.0 6.5

I

I

7.0

7.5

J

8.0

104/~. Fig. 2.-Clausius-Clapeyron plot for LizO(g) above LizO(s).

heat of sublimation then was obtained from an estimate of Aso,,b and the relationships AFQ = -RT In P

1248 1304 1327 1356 1384 1372 1300 1398 1447 1448 1467 1485 1534

8.013 7.669 7.536 7.375 7.331 7.289 7.194 7.153 6.911 6 ,906 6.817 6.734 6.519

I +Liz0 2.123 X 1.144 X 2.453 X 5.720 X 7.020 X

log PLi20

lo4 lo6 lo6

lo6 lo6 1.042 x 1 0 6 1.617 X lo6 1.980 x 106 7.810 X 108 7.590 X IO6

1.238 X lo7 1.771 X lo7 4.545 x 107

-9.530 -8.780 -8.451 -8.084 -7.020 -7.798 -7.602 -7.521 -6.900 -6.912 -6.694 -6.533 -6.110

Pmo

(atm.)

2.95 X 1.66 x 3.54 x 8.63 x 1.20 x 1.59 X 2.50 X 3.01 X 1.26 X 1.22 x 2.02 x 2.93 x 7.76 x

1O-Io

10-9 10-9 10-9 10-8 lo-* lo-'

lo-'

10-7 10-7 10-7

- TAXo

These values for the entropy and free energy funcIn a like manner, the heat a t absolute zero was then tion of the gaseous Li20, together with the observed pressure, yielded the values of AHTO and obtained from an estimate of A[(Fo - H O ~ ) / T ] ~absolute ,~ AHoowhich are presented in Table IV. The averand the relationship age values me AHTO = 9G.l kcal. mole-' and AHoO AH0 - AFo = 103.7 kcal. mole-I. The difference between A (Fa THO") T T the slope and third law values of AHoO is thus about The ion currents of &Of and the partial pressures 10.5%. Its most likely cause is temperature inhoof Liz0 gas are listed in Table I1 a t various tem- mogeneity within the Knudsen cell. peratures. TABLEI11 The entropy and free energy function of solid STATISTICAL ENTROPIES AND FREEENERGY FUNCTIONS Li20 were derived from the previously mentioned LiO Gas Liz0 Gas sources of data. However, an estimation of the - Fo - HoO -PO Hoo T SO T sa T entropy and free energy function of the gas necessitated the formulation of a molecular model and 70.384 59.551 50.301 58.126 1300 an estimation of vibrational frequencies. The in71.351 60.358 50.047 58.761 1400 ternuclear distance and stretching bond constant 72.263 61.123 51.488 50.356 1500 were chosen by comparison of LiO to the iso-elecD. The Equilibrium: LiO(g) = Li(g) tronic molecule BeF which is included in the compilation of molecular data by Herzberg.13 An l/zOz(g).--In the run with the isotopically enriched sample, the isotope ratio was 6Li/7Li = 0.825. 1. AF" = AH"

(OK.)

+

(13) G . Herzberg, "Moleculsr Spectra and Molecular Structure. Spectra of Diatomic Molecules," D. Van Nostrand Co., Ino., New York, N. Y . , 1.950.

(14) G. Herzberg, "Infra-Red and Ramen Spectra," D. Van Nostrand Co., Inc., New York, N. y . , 1945, pp. 172, 180.

. c

MASSSPECTROMETRIC STUDYOF

May, 1959

THE

647

SUBLIMATION OF LITHIUMOXIDE

TABLE IV STATISTICAL HEATSOF REACTION FOR LizO(s) = IlizO(g) AHDO T

AH@ (kcal.)

83.904 80.281 78.694 76.818 76.131 75.544 74.579 74.182 71.149 71.200 70.127 69.320 67.179

104.7 104.7 104.4 104.2 103.8 103.6 103.7 103.7 103.0 103.1 102.9 102.9 103.1

T (OK.)

ASTO

1248 1304 1327 1356 1364, 1372 1390 1398 1447 1448 1467 1485 1534

35.420 35.034 34.865 34.651 34.592 34.533 34.400 34.341 33.991 33.984 33.848 33.720 33.379

79.030 75.209 73.536 71.551 70.834 70.217 69.185 68.758 65.567 65.615 64.481 63.617 61.338

This was verified by monitoring the 6Li and 'Li peaks as well as the 7Liz'60 and 7Li6Li160peaks. The mass 28 peak could not be used due to the high background of carbon monoxide. Absolute pressures of LiO were then determined from the observed ion currents of 6Li'60+and combined with the free energy functions to yield the values of ANoO in Table V. The average value, AHo0 = 24.1 kcal. rnole-l, combined with the dissociation energy of oxygen, AGOo= 118.0 kcal. mole-', yielded the value BaO(Li0) = 3.6 e.v. for the dissociation energy of the LiO gaseous molecule. This is in qualitative agreement with the slope value Doo(LizO) = 7.0 e.v. or the 3rd law value Doo(Li20)= 6.5 e.v. since it is to be expected that the addition of one lithium atom to oxygen will release slightly more energy than the addition of a second lithium atom. The Doo(LiO) obtained must be considered only an upper limit, since the LiO+ was all assumed to be produced by simple ionization of the LiO molecule. The appearance potential of LiO (9.0 e.v.) is considerably higher than that of LizO+ (6.8 e.v.). If the observed LiO+ were produced by fragmentation of the Liz0 molecule, this value would indicate that the ionization potential of the LiO molecule is about 5.5-6.0 e.v. (assuming that AE for the reaction Liz0 + LiO Li is about 3-3.5 e.v., Le., slightly less than one-half the atomization energy of LizO).

+

TABLE V STATISTICAL HEATOF REACTION FOR

1346 1351 1390 1400 1444 1470

7.00 5.81 5.54 5.58 5.22 4.76

11.54 11.54 11.55 11.55

18.54 17.35 17.09 17.13

11.58

16.80

11.60

16.36

40.294 40.106 40.023 39.918 39.889 39.860 39.794 39.765 39.573 39.569 39.494 39.423 39.220

98.6 98. 1 97.6 97.0 96.6 96.3 96.2 96.1 94.9 95.0 94.6 94.5 94.1

tion of a second Li atom releases more energy than the first, which is unlikely. For the above reasons it appears quite probable that the observed LiO+ ion is a parent ion and that the value obtained for the dissociation energy of LiO is correct. TABLE VI COMPOSITION OF THE VAPORAT 1400°K. Li 0 2

Liz0 LiO

If the LiO+ were produced by fragmentation of LizO, the absence of an ion intensity of LiO+ below 9.0 e.v. enables one to estimate the maximum partial pressure of LiO that would evade detection. This partial pressure, about one-thirtieth of that shown in Table VI, would imply that Do(LiO) is less than 3.2 e.v. Since the atomization energy of Lip0 is about 7.0 e.v., this implies that the addi-

Mole % of vapor

x x x x

72.2 18.0 9.5 0.3

2.36 5.89 3.10 9.5

10-7 10-8 10-8 10-10

TABLE VI1 HEATSOF VARIOUSREACTIONS IN Reaction LizO(s) = LizO(g)

LipO(s) = 2Li(g) f '/zOz(g) LizO(g) = 2Li(g) Wg)

+ + +

* NOTEADDED IN

THE

Li-0 SYSTEM*

AHTO (kcal.)

AH140Oo = AHoo = AH14OO0 = AHoo = AH2gO0 = AH1400°

I,iO(g) = Li(g) '/zOz(g) LiO(g) = Li(g) O(g) 24.9 23.4 23.8 24.0 24.3 24.1

(atm.)

P

Species

+108.3 +115.9 96.1 +103.7 f219.3

+

= +214.6

$161.7

AH&'

=

AHoo

= +173.9

+ 24.1 = + 83.1

AHOO = AH00

Method

Exptl. slope Exptl. slope Exptl. 3rd law Exptl. 3rd law Available thermodynamic data Available thermodynamic data Availabledata and slope AHsuho Available data and 3rd law AHsubo

Experimental 3rd law Available data and 3rd law AH

PRooF.-since this paper was submitted for publication, some experimental data on the structure of the Liz0 molecule have become available. Electron diffraction studies by P. A. Akishin and N. G. Rambidi (Doklady A k a d . Nauk 8. p. S. R., 118, 973 (1958)) yield the value r(Li-0) = 1.82 A. and angle 110' for Li-0-Li. The use of these values in the calculation of thermodynamic functions results in an increase of S by 1.156 e.u. for LiO(g) and by 1.701 e x . for LizO(g). The correct thermodynamic functions, together with the absolute pressure data, give these corrected results:

648

K6z6 SHINODA, TERUKO YAMANAKA AND KYOJI KINOSHITA

Discussion The saturated vapor in equilibrium with solid Liz() contains gaseous Li, o2 and LizO in comparable amounts with LiO as a minor constituent. The composition of the vapor at 1 4 0 0 0 ~is. shown in Table Iv. In addition to the previous results, calculations using the limited data available for the LiONa (mass 46) peak show that ~ p = ~-5.2 ~ kcal./mole-l for the reaction Na(g) + LizO(g) = LiONa(g) + Li(g). The existence of detectable amounts of gaseous NazO molecules Over solid 80dium oxide is thus suggested. The heats of various reactions are given in Table VII. I n order t o see what effect the shape of the molecule has on the third law heats of reaction involving this molecule, a recalculation of the entrOpy a t 1400°K. was made assuming a linear model, but with the same bond constants. The resulting third law entropy for the linear model was 66.82 e.u.,

Vol. 63

compared with 71.35 e.u. for the bent model. At ,140OoK., the difference in AH would be TAS or 6.34 kcal. mole-l. Closer agreement between the slope and third law values of the heats is obtained USing thermodynamic quantities calculated for a bent model. This is to be expected since both covalent and ionic16models predict a bent form for the LizO molecule. In the case of the ionic model this bent ~ ~ 0 form is a result of the large polarizability of.the OXY$en ion and the high polarizing power of the lithium ion. The large amount of polarization in this molecule can also be considered as a large covalent contribution t o the bond. The third law heats are considered more reliable since the slope of the log P vs. 1/T curve could easily be too high if temperature gradients exist in the Knudsencell. (15) F. Hund, z. Physik, 32,1(1925).

STIEtFACE CHEMICAL PROPERTIES IN AQUEOUS SOLUTIONS OF ;PLTONIONIC SURFACTANTS: OCTYL GLYCOL ETHER, CY-OCTYLGLYCERYL ETHER AND OCTYL GLUCOSIDE BY K6z6 SHINODA, TERUKO YAMANAKA AND KYOJI KINOSHITA Department of Chemistry, Faculty of Engineemng, Yokohama National Universitu, Minamiku, Yokohama, Japan Received June $0, 1068

The surface tension, critical micelle concentration (c.m.c.), surface excess, foaminess and foam stability of aqueous solutions of octanol, octyl glycol ether, octyl glyceryl ether and octyl glucoside have been determined. The surface activity and/or c.m.c. values of non-ionic surfactants, containing the octyl group as the hydrocarbon chain, are similar to those of ionic surfactants containing the undecyl or dodecyl group as the hydrocarbon chain. As the foaminess and foam stability of the compounds in the series improved markedly with the increase in the size of the hydrophilic moiety, these properties are probably dependent upon the hydrophilic-lyophilic balance of the molecule.

Experimental

Introduction I n spite of the industrial importance of nonionic surface active agents, few reportsi-a concerning the surface chemical properties of the pure materials have been published, probably because the pure compounds are difficult to obtain; both the synthesis of pure polyoxyethylene alkyl ethers and the purification of commercial non-ionic surfactants are troublesome. We have investigated the surface chemical properties of a series of octyl poly01 ethers to determine the effect of differences in the hydrophilic group. The relatively short hydrocarbon chain, C,, was chosen as the hydrophobic group because (1) the c.m.c. values of these non-ionic surfactants are close to those of ionic surfactants containing the dodecyl group; (2) with Cs as the hydrocarbon chain, the hydrophilic-lyophilic balance changes considerably with an increase in the number of oxygen atoms from 1 to 6; and (3) the synthesis, purification and measurements of longer chain compounds are more difficult.

Materials.-Octanol obtained from the Ka6 Soap Co. Ltd. was purified by fractional distillation through a 100 cm. column to give a product boiling at 96' (16 mm.). Octyl glycol ether, synthesized4 from octyl bromide (b.p. 94-95" a t 20 mm.), and ethylene glycol (b.p. 116' at 40 mm.), wm purified by fractional distillation through a 60 cm. column, b.p. 132' (21 mm.); TPD 1.4355, n Z 0 D 1.4357.4 or-Octyl glyceryl ether prepared6 from glycerol a-monochlorohydrin b.p. 125-130' at 21 mm.) and sodium octylate, was puri ed by distillation, b.p. 132-133' (0.5 mm.); d20 0.9614; ~ 9 1.4517. 0 ~ 8-D-Octyl Glucoside.-Glucose was acetylated to give (3pentaacetyl glucose (m.p. 127.5-128.5') which upon bromination yielded acetobromoglucose, m.p. 87-88". This bromo compound reacted with octanol in the presence of silver oxide to give 8-tetraacetyl octyl glucosidesJ (m.p. 61.5-62.5'), which was deacetylated in sodium methylate solution in 24 hours a t 10-20' to yield P-octyl glucoside; m.p. 63.8-65'. (62-65',7 65-9906); [ C ~ ] ~ O D -33.8' in 4% aqueous solution ( [ a ] 2 0 ~ -34').7 Careful purification, drying and crystallization were indispensable in the synthesis of this compound. Procedures.-Surface tension was measured by the drop weight method, with a tip 0.249 cm. in diameter, in an air thermostat at 25 i 0.2". There was no appreciable change of surface tension with time within two minutes after the

(1) C. R. Bury and J. Browning, Trans. Faraday Soc., 49, 209 (1953). (2) T. Nakagawa, et al., J. Chem. SOC.Japan, 77, 1563 (1956); 79, 345. 348 (1958)(in Japanese). (3) L. M. Kushner, W. D. Hubbard and A. S. Doan, THISJOURNAL, 61, 371 (1957).

(4) F. C. Cooper and M. W. Partridge, J. Chem. Soc., 459 (1950). (5) G. G. Davies and W. M. Owens, ibid., 132, 2542 (1930). (6) C. R. Noller and W. C. Rookwell, J. A m . Ckem. SOC.,60, 2076 (1938). (7) W. W. Pigman and N. K. Riohtmyer, ibid., 64, 369 (1942).

x