December 15, 1923] 



NA TURE 



869 



If this be the case, the closer packing or compression 

 if the juxtaposed molecules in the unimolecular films, 

 IS revealed in the investigations of Devaux, Langmuir, 

 ind Adam, may be to some extent explained by the 

 straightening out of these zig-zags, or perhaps by the 

 ■' elastic compression " of the helices. 



As pointed out by Langmuir, the question of the 

 formation of unimolecular surface films can be attacked 

 in a different manner. It is known that gases or 

 \apours can be condensed or adsorbed by solid and 

 liquid surfaces. The question then arises, does the 

 lormation of primar}' unimolecular films ever occur in 

 such cases ? It will be recollected that Hardy made the 

 suggestion that the formation of the primary uni- 

 molecular film in the spreading of oily substances on 

 water might be due to adsorption from the vapour. In 

 order to examine this question, Mr. T. Iredale has 

 recently measured in my laboratorj^ the fall in the 

 surface tension of mercury caused by exposing a fresh 

 mercury surface to vapours of increasing partial 

 pressure. The excess surface concentration q of the 

 adsorbed vapour can then be calculated by means of 

 Gibbs's formula 



where y = surface tension, and p and p denote the 

 density and partial pressure of the vapour respectively. 

 Working with the vapour of methyl acetate, Iredale 

 found in this way that at a temperature of 26° C. and a 

 partial pressure of 62 mm. of mercury, ^ = 4-5 x 10"® gm. 

 per square centimetre of surface. From this result 

 we can readily calculate that there are 0-37 x 10^^ 

 molecules of methyl acetate adsorbed per sq. cm., 

 and that the area per molecule is 27 x io~^^ sq. cm. 

 As under the conditions corresponding to this calcula- 

 tion the molecular surface layer was probably not 

 (juite saturated (in the unimolecular sense), we may 

 expect the value found to be of the same order of 

 magnitude but somewhat greater than the values 

 found by Adam for the cross section of the head group 

 of the higher saturated fatty acids (25 x 10 i^) and of 

 the esters (22x10^1* for ethyl palmitate and ethyl 

 behenate). We may, therefore, say that Iredale's 

 results appear to indicate the formation of a primary 

 unimolecular layer built up by adsorption from the 

 vapour phase. 



Langmuir has measured the adsorption of a number 



if gases at low temperatures and pressures on measured 



urfaces of mica and glass, and has arrived at the 



inclusion that the maximum quantities adsorbed are 



always somewhat less than the amounts to be expected 



in a unimolecular surface layer, E. K. Carver, who has 



measured the adsorption of toluene vapour on known 



lass surfaces, has arrived at a similar conclusion. The 



\ lew that the maximum adsorption from the gas phase 



( unnot exceed a unimolecular layer has, however, been 



much criticised. 



Let us now consider another type of formation of 

 urface layers at the surfaces of liquids— namely, the 

 ■ .ise where a substance dissolved in a liquid concentrates 

 preferentially at the liquid-air or liquid-vapour inter- 

 face. Gil)bs, and later J. J. Thomson, have shown 

 that if a dissolved substance (in relatively dilute 

 solution) lowers the surface tension, it will concentrate 



NO. 2824, VOL. I 12] 



at the surface. That such a phenomenon actually 

 occurs has been qualitatively demonstrated in the 

 experiments of D. H. Hall, J. von Zawidski, and 

 F. B. Kenrick and C. Benson, by the analysis of foams 

 and froths. In 1908 S. R. Milner used the same 

 method in the case of aqueous solutions of sodium 

 oleate, and arrived at a mean value of i'2 x 10'"^'' gram 

 mols. excess concentration per sq. cm. of surface. 

 In the case of dilute solution, we can calculate q, the 

 amount concentrated or " adsorbed " in the surface 

 per sq. cm. (excess surface concentration), and Milner 

 calculated from Whatmough's data for aqueous 

 solutions of acetic acid that the " saturation " value 

 of q is 3*3 X io~^" mols. per sq. cm., from which it 

 follows that the area per molecule in the surface is 

 50 X 10^1^ sq. cm. In a similar manner, Langmuir has 

 calculated from B. de Szyszkowski's data for aqueous 

 solutions of propionic, butyric, valeric, and caproic 

 acids that the surface area per molecule adsorbed in 

 the saturated layer is equal to 31 x lo"^® sq. cm., while 

 Harkins has arrived from his own measurements for 

 butyric acid at the value 36 x 10"^^ sq. cm. 



In 191 1 Dr. J. T. Barker and I made a direct 

 determination of q for a solution of nonylic acid in 

 water. For a practically saturated surface layer it 

 was found that q was about i-o x 10"'' grm. per sq. cm., 

 or 3-1 X iqI* molecules per sq. cm. From this result 

 it follows that the surface area per molecule is 

 26 X io~^^ sq. cm. 



These values are not ver}' different from the 

 values found by Langmuir and by Adam for the 

 oriented unimolecular layers of practically insoluble 

 fatty acids resting on the surface of water. That in 

 the present case some of the values are larger might 

 easily be explained on the ground that these adsorption 

 layers are partially, or completely, in the state of 

 " surface vapours." For Adam and Marcelin have 

 recently made the important discovery that the 

 unimolecular surface films investigated by them may 

 pass rapidly on increase of temperature from the 

 state of " solid " or " Uquid " surface films to the state 

 of " vaporised " surface films, in which the juxtaposed 

 molecules become detached from each other and move 

 about with a Brownian or quasi-molecular motion. 



It is, indeed, highly probable that the molecules 

 which are concentrated in the surface from the state 

 of solution in the liquid phase are not in quite the 

 same situation as the molecules of practically insoluble 

 substances which are placed on the surface. In the 

 former case the molecules are still " dissolved," so that 

 they will be more subject to thermal agitation and 

 less able to form a juxtaposed unimolecular layer. 

 They may also be " hydrated." Nevertheless, the 

 agreement as regards order of magnitude in the values 

 of the surface area per molecule in the two types of 

 case is certainly very suggestive and significant. 



Let me now direct attention to another very 

 interesting phenomenon relating to the surfaces of 

 liquids and solutions — namely, the existence of an 

 electrical potential gradient or potential difference 

 (P.D.) in the surface layer. The liquid-gas interface 

 offers the simplest case of such interfaces, so 

 the investigation of the potential differences which 

 may exist at this interface is a matter of funda- 

 mental interest. In 1896 F. B. Kenrick developed, 



