NOTES
762
+
time-lag to be multiplied by ( V , V,>/V,. This is confirmed by the calculated results shown in F’g ‘ 1 ure 1.
In Fig. 1 the transient part of curve 4 has not yet reached the asymptote a t Dt/e2L2 = 1.0. An examination of the calculated data beyond the range of Fig. 1 showed that the linear asymptote was reached at Dt/e2L2= 4.0 and that the intercept of this line on the abscissa is -0.494, in agreement with eq. 5 and as shown in Fig. 1for the values of V I ,VZ,and H chosen above. Acknowledgment.-The author wishes to thank the donors of the Petroleum Research Fund, administered by the American Chemical Society, for their support of the research which led to this paper.
Vol. 66
liquid, r = N A / ~2 ,is the external area of the solid, and ANI, the molar heat of liquefaction. The net integral heat of adsorption hnet = ( J Z ~ ( ~ L ) JzI(sfL)) is used most frequently and hr values are expressed commonly in units of ergs/cm.2 of surface. The need to define hL, the enthalpy change involved in destroying a cm.2 of liquid surface, will become apparent. The difference in entropy of the adsorbed film, SA,and the entropy of the bulk liquid is given by the expression T ( S s - SL)=
[ ( ~ I ( s L )-
hI(sfL)/rl
+ p/r - k T l n X
(2)
Here X(= PIPo) is the relative equilibrium pressure and p (= YSD - ysf) the difference in surface free energy of the absorbate-free solid and one containing adsorbed molecules at a surface concentration r (expressed best in moles/cm.2 to yield conA CRITICISM OF HEAT OF IMMERSIONAL ventional entropy units). The equation is valid also if the solid is perturbed on adsorption; the WETTING PRACTICE symbol SA’is used to denote the entropy of the BYJ. J. CHESSICK adsorbate-solid system now. 2. The Use of Limited Heat of Wetting Data Wzlliam H. Chandler Chemistry Laboratory, Lehigh Universzty, Bethlehem, Penna. For the Interpretation of Solid-Liquid Interactions. Recezved September 81, 1061 --Many workers in the past have reported and used The procedural techniques of heat of wetting for interpretive purposes only heat values hI(SL) practice as well as those of data interpretation for the immersion of a given solid into a variety of used by workers in the 1930’s generally persist liquids or different solids into one liquid. Signito this day. Some modifications have been in- ficantly, the writer is co-author of two of these troduced, obviously, but these for the most part are articles. The following references are four of either correctional in nature or due t o advances many.2 Recently, a paper pointing out a few of in equipment. The question raised here is whether the dangers inherent in the use of insufficient data more than modest changes are needed. Evidence was published.a Here principal attention was paid is presented to support the view that, indeed, drastic to reported ~ I ( S L )values for the immersion of a changes are needed to prevent errors of experiment rutile (TiOz) temperature activated a t 500°, and interpretations of inadequate data in much of into a variety of n-polar paraffin derivatives (three and four carbon compounds such as the acid, the current writing in this field. The number of articles which in the opinion of amine, alcohol, halides) of references 2a and 2b. the writer could be included here is large. The There the net heats of adsorption (hI(sL) - h(L))/F requirement of brevity alone is sufficient justi- a t monolayer coverage were compared with calfication for the restriction of discussion to only a culated interaction energies; the sum of solidlimited number of authors. This, of course, i s adsorbate energies due to dipole, polarization, unfair provided the reader restricts his thinking and dispersion, and adsorbate-adsorbate interaction judgr ent to the articles discussed here. A good affects. Further, t,he dope of the linear curve portion of the paper treats critically articles co- obtained when the net integral energies of interauthored by the writer. The paper suffers as a action were plotted against the dipole moments of consequence since in many instances better argu- the wetting liquids was used to estimate the average force field emanating from the solid surface. ments could have been presented. The calculations of net integral heats of adsorpDiscussion tion of n-butyl derivatives onto rutile2as2bfrom ) alone required the following assump1. The Relationship between Heat of Immer- hI ( 8 ~ values tions: (1) the formation of a monolayer of wetting sional Wetting and Adsorption Thermodynamics.Only a brief discussion of basic relationships, back- liquid molecules oriented perpendicular to the ground material for the discussion which follows, surface, polar groups down, (2) the same number will be given. A detailed treatment has been of close-packed adsorbate molecules per unit area of surface for the seven different n-paraffin derivapresented. The integral heat of adsorption of N A moles of tives; (3) a value of hI(SfL) at e = 1 equal to hL adsorbate from the vapor state at saturation pres- for the particular wetting liquid; (4) packing independent of the substrate; (5) physical adsorpsure POand temperature T is tion alone occurred on high temperature activated hnda. = [(hI(BL) - hI(SfL))/rl -t AHL (1) (2) (a) F. H. Healey, J. J. Chessiok, G. J. Young, and A. C. Zettlewhere hI(SL) and ~ I ( s ~ Lare, ) respectively, the heat moyer, J . Phys. Chem., 68, 887 (1954); (b) J. J. Chessiok, A. C. Zettleevolved on immersing an adsorbate-free (evacuated) moyer, F. H. Healey, and G. J. Young, Can. J . Chem., 83, 251 (1955); (c) E. V. Il’in and V. F. Kiselev, Doklady Akad. Nauk, S.S.S.R., 82, sample and a sample containing NA moles of pre(1952); (d) W. D. Harkins, “The Physical Chemistry of Surface adsorbed molecules of wetting liquid into this 85 Films,” Reinhold Publ. Corp., New York, N. Y.,1952. ( 1 ) J. J. Chessick and 263-9 (1959).
$. C. Zettlemoyer,
Adz’ances qn C a t a l p i s , 11,
(3) C. M. Hollabaugh and J. J. Chessick, J . Pkvs. Chem., 66, 109 (1961).
April, 1962
NOTES
763
Further the assumption that hI(SfL)a t 0 = 1 is rutile and others. The work of Hollabaugh and Chessick3 demonstrated clearly that some (and equal to hL for the wetting liquid can be seriously undoubtedly all) of these assumptions are incorrect. incorrect when organic immersion liquids are emFor example, n-butyl halides do not absorb to form ployed. The heats of immersion of rutile into noriented monolayers; rather, fiat-wise adsorption but)yl chloride or n-propyl alcohol measured as a as well as multilayer formation was observed. function of pre-adsorbed wetting liquid revealed The predicted orientation of alcohol molecules did that heats evolved per cm.2 a t monolayer coveroccur but the packing was not independent of the ages are markedly different than the respective substrate and chemisorption could not be ruled hL values.3 Similar findings were reported for out. Other workerszc made calculations of this graphite samples immersed in toluene, carbon tetrachloride, n-heptane, cyclohexane, and ntype using inadequate data also. Obviously, the use of plots of ( ~ I ( s L ) - hr(sfL)) propyl alcohol.’ The paucity in the literature of us. dipole moment to calculate the average field heat of immersion curves obtained as a function of 0 emanating from a solid surface must be viewed for solid-organic liquid systems, which could with suspicion. Despite the warning given3against justify the use of reasonable, assumed values of this practice, such a calculation from too limited hI(sfL)at 0 = 1or greater in certain instances, is an experimental data was carried out r e ~ e n t l y . ~additional reason to refrain from the uncertain and A serious objection, purely experimental in nature, carry out the necessary, more tedious chore of obcan be advanced here. The workers of references taining isotherm data and heats of wetting as a func2a and 4 used n-butyl chloride as an immersion tion of the amounts preadsorbed. The paucity liquid because of its high dipole moment relative of such data is convincing evidence that the easy to the hydrocarbon, amine, alcohol, etc., completely path of limited experiment and unproved assumpignoring the fact that an alcohol and amine pos- tions is stili with us. 4. The Use of h~ Values to Calculate Net sess a peripheral, positive dipole and the chloride a non-peripheral, negative dipole. Their measured Integral Heats of Adsorption : Solid-Water Systems.-The collection and inspection of the more ~ I ( S L )values were large, nevertheless, and in apparent agreement with the predicted increase in numerous heat of wetting values obtained as a heat evolved per unit area of solid with increase in function of e for the immersion of non-porous dipole moment ol the wetting liquid. These values polar solids having surface areas ca. 5 to 20 m.”/g. are experimentally incorrect and result from the in water has shown that rarely does a hI(sfL) presence of unsuspected trace amounts of water in value exceed hL by more than 20% a t e = 1. At the wetting liquid. Employment of the multi-bulb higher coverages the use of hL rather than the technique developed by Wightman5 to remove this measured value has a good measure of support, trace water reduced the heat of immersion of ru- provided a duplex film forms. Such an assumptile in n-butyl chloride from 621 21 to 281 f 12 tion need not be made, however; film type can be determined experimentally. ergs/~m.~. Heat of immersion values, ~ I ( s L ) , less than h~ An understanding of the adsorption process demands isotherm data, preferably measured rather are indicative of systems upon which non-duplex than assumed data, over the range of relative pres- films form. Nevertheless, net integral heats - ILL) have been calculated for such systems, sure from zero to that of the coverage of interest, (~I(sL) as well as heat ad immersion values obtained as a e.g., graphite6c and silica6bin water. The reason function of the amount of wetting liquid pre-ad- for a negative net integral heat of adsorption sorbed before immersion. This is a long and [ ( ~ I ( B L )- h ~ =) (108 - 118) = -10 ergs/cm.Z)] tedious process even for one adsorbent-absorbate found in this Laboratory for the immersion of the silica Aerosil (now Cab-0-Sil) was revealed by systems. 3. The Use of h~ Values to Calculate Net further water and nitrogen isotherm determiIntegral Heats of Adsorption ; Solid-Organic Liquid nations. These showed that the silica surface was Systems.-Where net integral heats of adsorption 75% hydrophobic (non-duplex film) and 25% (~I(sL) - hr(sfL)), expressed here in units of ergs/ hydrophilic (possibly duplex film formation if cm.,2 are used for interpretative purposes, most patches large enough). A similar, fully hydroxyfrequently, h ~ the , heat evolved upon destruction lated, 25’ outgassed silica had a ~ I ( S L )value of 435 of a unit area of (wetting) liquid surface is used in erg~/cm.~. It is true that the quantity ( ~ I ( s L ) - h ~ can ) place of hI(sfL)at 0 = 1 or greater.* While this practice may not always lead to serious difficulty, be used simply as a comparative device. This must particularly if employed with ~ I ( S L ) values for be plainly stated, however, when this procedure is polar solid-water systems, it can be condemned employed. Perhaps it would be more convincing on grounds that assumed values are unnecessary. to state that the quantity (F.I(sL) - h I f s f L ) ) has The appropriate hT(sfL) values can be obtained ex- thermodynamic significance, irrespective of whether the adsorption is non-duplex or duplex. This is perimentally without difficulty. illustrated by the established equation L. Romo, J . Colloid Sei., 16, 139 (1961). (8) J. P. Wightman, Ph.D. Thesis, Lehigh University, Bethlehem,
(4)
Pa., 1960. (6) (a) F. E. Bartell, J. Phys. Chem., 68, 36 (1954); (b) D. J. Hine and W. C . Wake, Trans. Faraday Soc., 66, 1017 (1959); (c) R. L. Every, W. E. Wade, and N. Hackerman, J . Phys. Chem., 66, 25 (1961); fd) See previous references 2a and 2b also. Many others are unlisted.
-
( ~ I ~ s L ) hI(sfL))
=
(EA‘- E L ) = qat d r
- rAHL
(3)
Only E’A, the molar energy of the adsorbate-solid (7) J. %J.Chessick, A. C. Zettlemoyer, and Yung-F’rtnp Pu,J . Phgs. Chern., 64, 530 (1960).
764
NOTES
system, and EL, the molar energy of the liquid adsorbate, have not been defined previously. The quantity ( ~ I ( s L ) - h ~ has ) thermodynamic sigriificance only if hI(sfL)= hL. 5. Heats of Immersion of High-Temperature Activated Solids.--Almost never are solids evacuated at elevated temperatures (300 to 500") prior to immersion on conventional apparatus protected from the influence of trace amounts of organic contaminants in the system. This statement applies to all references of this paper except 1 (in part), 3, 5 , and 7 although it is possible that in some (limited) instances precautions were taken but not reported. Heat values (ergs/cm.2) measured for the immersion of rutile evacuated at 450' to an ultimate vacuum of 10" mm. on a B.E.T.-type adsorption rig were much too erratic to be acceptable.3 The heat of immersion of rutile outgassed in this manner in water was found to be 708 i 20 ergs/ cm.2. Very divergent values were rejected on a statistical basis. If a liquid nitrogen trap separated the sample from the adsorption system during evacuation and further if dry oxygen was introduced in the sample system at the evacuation temperature, the heat of immersion of this rutile sample decreased to 588 i 9 ergs/cm.2. Reduction of the rutile surface is indicated; organics are present on the solid surface and (or) on the rig. Similar divergent heat values were found recently in this Laboratory for the immersion of protected and unprotected samples of alumina activated at 500'. In this instance, the use of the nitrogen trap alone was sufficient to prevent surface contamination; unlike the rutile sample its surface originally was free of organic contaminants. Footnote 10 of a recent publication which reported the heats of immersion, h I ( S I , ) , of a variety of rutile and anatase samples in water as a functiop of temperature of activation* reveals a belated recognition of this unsuspected, additional variable. Most articles, reporting immersion values for temperature-activated solids, suffer from this defect. The assumption made too frequently that polar solids such as anatase,Zd rutile,2aJh calcium or barium sulfate,2o and many others activated at temperatures near 500' only physically adsorb molecules of the wetting liquid can not be accepted without further detailed studies.a There is danger also in comparing ( h ~ ( s L )- ~ I ( L ) ) values with spreading pressures, p, correctly calculated to e's at the arbitrarily selected relative equilibrium pressure, both quantities being expressed and compared in units of ergs/cm.2.9 Free energies, enthalpies, and, particularly, entropies expressed in units of ergs/cm.2 or ergs/ cm.2/0C are much more difficult to interpret than those expressed in conventional units. A further need for converting cp and hnet values into conventional units is seen from inspection of eq. 2. Spreading pressure values are accurately comparable only to net integral heats when corrected by t'he RT In PIPo term; this was not done.g In addition, com(8) T. H. Wade and N. Hackerman, J . Phgs. Chem.. 66, 1861 (1961). (9) R. L. Every, W. H. Wade, and N. Hackermen, ibid., 66, 25 (1961).
VoI. 66
parison of integral heat, free energy, and entropy values at a relative pressure so high that multilayers form suffers because of the insensitivity of such values to surface-adsorbate effects so pronounced in the region zero to monolayer coverage. Capillary effects due to interparticle or particle pores and as a consequence the decrease in area of the solid sample offer further possible complications. Acknowledgment.-The author is grateful to the U. S. Air Force, Wright Air Development Division, Materials Center, Wright-Patterson Air Force Base, Ohio and the National Printing Ink Research Institute, Lehigh University, Bethlehem, Pennsylvania, for support in the development of this paper. FLUORINE N.M.R. SPECTROSCOPY. VI. FLUOROCARBON SULFIDES' BYGEORGEVAN DYKETIERS Contribution No. 214 from the Central Research Dept., Minnesota Mining & M f g . Co., St. Paul 19, Mznn. Received October 21, 1981
While nuclear magnetic resonance (n.m.r.) data have been reported for several fluorocarbon derivatives of sulfur hexafluoride12~3 and for certain unusual fluorocarbon s~lfides,*~6 such data for simple fluorocarbon mono- or polysulfides have been lacking. The n.m.r. measurements here presented fill this gap; and owing to their high precision several significant generalizations concerning fluorine n.m.r. spectroscopy also emerge from this work. Experimental The syntheses and physical properties of the fluorocarbon sulfides have been reported,e-B save for the ultraviolet spectrum and certain constants of 1,4-dithiaperfluorocy~10hexane. This compound, b.p. 81.5", f.p. -6.5", was studied a t a concentration of 1 pl./ml. in isooctane, and had X., = 2232 A. and ernax. = 132. From its refractive index, 12% 1.3585, (lit. 1.35688) and density, d26 1.693, there is calculated M R = 34.28, from which the atomic refractivity 7.40 is obtained for sulfur, in excellent agreement with the value previously given for fluorinated thmethers.? The n.m.r. equipment and techniques have been described.s In Table I are presented numerical shielding values, on.the +scale,g for all fluorine positions in each compound studied. Procedures for the approximate conversion of older and less precise numerical dataa-6 into +values have been given.s
Discussion ShieIding values for the -CF2S- group are seen to fall generally in the region 80-95 +, about midway between +-values for the isoelectronic P and C1 compounds.l0 Much narrower limits are appropriate when similar structures are compared. (I) Part I11 of the latter series; presented a t the 138th National A.C.S. Meeting, New York, 1960. (2) N. Muller, P.C. Lauterbur and G. F, Svatos. J . A m . Chem. 8 0 C . , 79, 1043 (1957). (3) R. P. Dresdner and J. A. Young, ibid.. 81,574 (1959). (4) M. Hauptsohein and M. Braid, %bid., 80,853 (1958). ( 5 ) C. G.Krespan, ibid., 83, 3434 (1961). (6) G.V. D. Tiers, J . Org. Chem., 26, 3515 (1961). (7) G.V. D. Tiers, ibid., 26, 2538 (1961). (8) C. G. Krespan, U.S. Pat. 2,931,803(April 5, 1960). (9) G.Filipovich and G. V. D. Tiers, J. Phys. Chem., 63,761 (1959). (10) E. Pitcher, A. D. Buokingham and F. G. A. Stone, J. Chem. Phys., 3 6 , 124 (1962).