33477
1 ) i i . w 'I'R.\NSITIONS IN SURFACE FiLais ON SOLIDS ~ ~ O S T R I H C . l ' I O FROM N T H E D E P A R T M E N T OF C H E M I S T R Y , POMONA C V L L E G E ]
Phase Transitions in Surface Films on Solids' B Y K . NELSONSMITH RECEIVED AVGUST21, 1951
Published isotherins of rt-heptane 011 Fe.03 a i i d graphite have !been redetermined usiug a gravimetric niethod and McLeod gage instead of a volumetric method and cathetometer. T h e present results are a t variance with the others in that no firstor second-order phase transitions are obser\.etl
positive and negative net heats of adsorption over a narrow temperature range. This behavior is contrary to general experience, as the authors point out. Further, the molecular areas calculated for the adsorbate held in these monolayers are enormously large, thus indicating a phase transition where but a fraction of the surface is covered, For these reasons and because of the importance of these phase transitions in theoretical interpretations of adsorption and of the nature of surface films, it was decided to repeat some of the isotherms e' = = r , - g = 1' , * - .{.! = f < Y ' ~f others by a different method. The two systems where p, etc., is the spreading pressure a n d I' is the which we felt had beqn in\-estigated in greatest desurface concentration in excess of that ill the gas tail, n-heptane on ferric oxide and n-heptane on phase. Using these computed spreading pressures graphite, have been studied again with great care a t GreggI3Harkins and associate^^-^ and othersI0-l2 \yarious temperatures. No evidence of a phase have interpreted films on solids by drawing analo- transition was found ior either system. gies to the known properties of insoluble films on Experimental liquids. The It-heptarie used these measurements was purThe strongest direct evidence in support of the chased froin the SationalforBureau of Standards with the imview that films on solid surfaces behave like insol- purity given as 0.1 niole per cent. It was stored over outuble films on liquids is the existence of discontinui- gassed metallic sodium for several months before distillation ties in experimental adsorption isotherms when the in oacuo to its reservoir in the vacuum line. The FsOI Baker and Adamson, reagent grade, with the maxifraction of the surface covered is less than unity, was mum limits of impurities listed as 0.25% insoluble in HCI, Benton and White,l8Ross and Boyd,14and Harkins, 0.25% SO4, 0.005% Cu, 0.005% Zn and 0.20% substances Jura and L o e ~ e r , ~have - ~ reported isotherms show- uot precipitated by S H 4 0 H . The graphite sample was ing such discontinuities. In most cases the samples obtained from the laboratory of W. D . Harkins and was thought to be like the sample for which Harkitis, Jura and were of a type that one would not expect t o possess Loeser report their phase transitions. uniform surfaces and consequently the occurrence The degassing procedure of Harkiris, Jura atid Loescre (Jf first order phase transitions is difficult to explain. \viis a t first followed, namely, evacuation at room temperature O 11 hours a t On re-examination of this work it is noted that for several hours followed by heating in V ~ U for ,725" and subsequent cooling in ' ~ " c z ~ o . The samples were for most of the n-heptane cases cited by Harkins, open t o continuous pumping during all stages. I t was found Jura and Loeser the phase transitions occur a t that this procedure did not satisfactorily outgas the samples, about the same absolute pressure, regardless of the for the pressure build-up in our vacuum system was about mm. adsorbent used or the temperature. I t is also noted 5 X 10-6 mtn. per hour for Feros and about 15 X that the isotherms for different temperatures cross per hour for the graphite over a two-day period compared to a pressurc build-up of 1.6 X mni. per hour for the one another in such a manner as to indicate both suine system without the samples. The adsorbent samples
In 1937 Banghani' showed that the Gibbs adsorption equation could be applied to the adsorption of vapors at solid surfaces and that the reduction in free surface energy can be identified with a tangential spreading. pressure of a surface film. Such a spreading pressure of an adsorbed film on a solid surface cannot be measured directly, as it can for a f i l ~ i on i a liquid surface, but it can be calculated by use of the equation (with symbols employed by various authors! ^I
L' :;
( 1 ) This is a progress report of work done under Contrdct 5 8 ONR-.;4700 with the Ofice of Naval Research. t2) D. H. Bangham, Trans. F a r a d a y S u r . , 33, SO5 (1937). f : i J S. J, Gregg, J . Cheni. Soc., 696 (lW2). 4) G. Jura and W. D. Harkins, THISJ O C K S A L , 66, 1356 (1944). 3) G . Jura, E. H . Loeser, P. R. Basford and 1%'. D. Harkins, J . i I i r i i i . P h y s . , 13, 535 (1945). (6) W. D. Harkins, G. Jura and I < . H . Loeser, THISJ O U R N A L , 68, .!.j4 (1946). (7) G. Jura, E. €I. Loeser, P. 11. Hnsford a n d W. D. Harkins, J . < ' / E M . Phys., 14, 117 (1916). 18) G. Jura, W. D. Harkins and E. H. Loeser, ibid., 14, 344 (1946). 19) J. Alereader, "Colloid Chemidtry," Vol. VI, Reinhold Publishing Corp., New York, N. Y . , 1946, chapter 1 by W. D. Herklns and G.Jura. (10) W. 0.Palmer, Proc. Roy. SOC.( L o h d o n ) , 8186, 264 (1941). (11) M. H. A. Armbrtlstet and J. B. Austin, THISJOURNAL, 66, I59 (1844). (12) M. A. Cook aid 1) H. Pack, i b i d , T i , 791 (1940). (13) A. F. Benton and T. A, White, ibiJ., 63, 3301 (1931). (14) S. Ross and G. E. R o y d , Document .MDDC-864, U.S. Atomic l'iiergy Commission, O a k Riclge, Tennessee. Declassified April 11, l!li7,
finally used were exposed several times t o n-he tane vapor and outgassed in oacuo overnight a t about 500 between exposures. The total heated degassing time for each sample \vas well over 50 hours. At the completion of each isotherm, the adsorbent was outgassed overnight at 200' and its weight checked to ascertain whether permanent weight changes had occurred during the course of a series of measurements. These weights always checked to within 0.3 mg. of the original weight, an amount which corresponds to 2.9 X cc. (S.T.P.) per gram for Fez08 and 2.0 X cc. for graphite. The graphite sample weighed 33.5 g. and the Fez08 weighed 23.1 g. The FenOs was not really outgassed by this procedure, for by putting a sample of FsOa into a quartz vessel and heating it t o successively higher temperatures, successively greater weight losses we.re observed accompanied by sintering and a great reduction in bulk volume, An isotherm run on the sintered sample showed considerably less adsorption than for the sample given the usual treatment. The treatment described does, however, give a satisfactory low pressure and the weight remains constant. The vacuum system included a medium-pressure McLeod gage, two reservoirs for n-heptane, precision-ground hollow as
t
l for the difference between our results and those of Harkins, Jura and Loeser. I t may be that our samples are not the same as those used by them (we did use the same brand and grade). The temperatures of our baths seem to be regulated with the same tolerance, and the n-heptane seems to be ot the same high quality. Our techniques do difier in three major ways A s previously described, the outgassing procedurc used by them was not adequate for us. Instead of the volumetric method, we used d gravimetric method which is free of the potential danger of accuniulating foreign or displaced gases. This may be very important a t very low pressures, especially if there is a tendency for the sample to build up pressure o f its own accord. In the gravimetric method the whole system except the sample bulb itself is highly evacuated between each isotherm point; this would tend to minimize such an d e c t . Finally, whereas we used a McLeod gage for pressure measurements, they used a manometer “with a traveling microscope of a sensitivity of 0.001 mm., which was set on a transit mount so that it could be rotated. The average deviation of a given height when a series of readings were taken was found to be 0.002 to 0.003 mrn., while that of a single reading from the average did not exceed 0.006 mm. The screw of the nlicroscope was not in error by niorc thdn 0.01 r i m . , and its errors are known to 0.001 Inni.”4 111our measurements, the observation of a 0.1 m u . pressure difference corresponded to 1.31 X 10 iniii., or better than A tenfold increase in sensitivity over their method. 111this C O I I I I ~ C ~ it~ ~I I LI LI V be noted that Whytl,Lw-Gr,iy ant1 Teich, l X ,tfteI making a detailed study of the mercury meniscus and its bearing on the precision measurements of gas pressures, conclude that “a precision of f0.01 rnm in the pressure reading is about the limit of accuracy that can be reached with a good cathetometer and scale.” Their discussion leads one to the conclusion that a cathetometer is not a suitable instrum n t for rnrdsureinent of pressures as low clone 118) K Whytla.ir G r i r and .U I eich, T r o w I mads) S L ) ~4 4 , 7 7 4 ( 1 ’) IR)
PIIASE TRANSITIONS IN SURFACE FILMSON SOLIDS
July 20, 1932
3479
I
.3P
I
P, mm.
Fig. 1.-Low pressure adsorption isotherms of n-heptane on graphite: 0, 21.1'; 0 , 27.6'; @, 34.3': 0, 41.0'. Dotted line is 25.0' isotherm of Harkins, Jura and Loeser on similar sample. I
I
,321
.28
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
2
3 P, mn.
4
Fig. 3.-Medium pressure adsorption isotherms of nheptane on graphite: 0, 21.1'; 0 , 27.6'; @, 34.3': 0 , 41.0'.
I"'""""'"""""I
I 2
Fig. 2.-Low pressure adsorption isotherms of n-heptane on Fe20a: 0, 8.1'; 0 , 14.6'; (3, 21 1'; 0 , 27.6'. Dotted line is 22.0' isotherm of Harkins, Jura and Loeser for a similar sample.
P. mm.
3
Fig. 4.-Medium pressure adsorption isotherms of n-heptane on Fe20t: 0, 8.1'; 0 , 14.6'; @, 21.1'; 0 , 27.6".
liquid phase increases as less active sites become effective. Thus, up to the formation of a monolayer there has been a continuous phase transition, but for these adsorptions. None of these factors, how- unlike the common transition phenomena i t has ocever, gives a satisfactory explanation for the ob- curred over a range of pressures. In special inserved difference. stances, solid surfaces may exist which do not have a continuous range of active sites, as in the case of Discussion The lack of any discontinuity in our n-heptane the specially prepared crystals of sodium and potasisotherms for ferric oxide and graphite establishes a sium chloride used by Ross and Boyd14for the adFor such a surface, reasonable doubt as to the reality of phase transi- sorption of ethane a t -183'. tions on heterogeneous surfaces when the fraction of one might observe a discontinuity in the adsorption the surface covered is small. We believe that the isotherm, not because of the occurrence of a special present status of these phase phenomena can be im- phase of the surface film, but because the existence proved only by further investigation by others for of a large area of relatively uniform surface activity various systems, using a variety of improved pres- permits the normal process of adsorption to occur a t a given pressure instead of over a range of pressure and adsorption techniques. A logical general concept of adsorption on hetero- sures. Further, if a large number of discontinuities geneous surfaces might consider the surface film to should exist in the same experimental isotherm, i t be composed of two phases-two-dimensional liquid would seem more logical to attribute the cause to islands connected by a two-dimensional gas. The surface heterogeneity of this type rather than to a liquid islands exist on the more active adsorption large variety of surface film phases. sites of the solid, and as the pressure increases the CLAREMONT,CALIF.
