Desorption of Electrolytes at Liquid-Vapour andLiquid-Liquid InterfacesBY ROBERT AVEYARD," SYED M. SALEEM AND (IN PART) ROY HESELDENDepartment of Chemistry, The University of Hull, Hull HU6 7RXReceived 28th June, I976The desorption of various simple inorganic salts at the decanol-water interface has been studied,and analysed together with previously obtained results for the air-electrolyte and dodecane-electrolyte interfaces. The desorption of the salts is much more dependent on the nature of theanion, at all three interface types, than it is on the cation and the results are discussed in terms of thepossible interaction between anions and hydration layers at the interfaces. The dependence of freeenergies of adsorption of alkanols, both from alkane and from electrolyte, on the alkane-electrolyteinterface has also been investigated and it is concluded that the chemical potential of the alkanolat the surface is not much affected by the salts, but the influence is greater the less the desorption ofthe salt.In a recent paper we considered the effect of simple inorganic electrolytes on theinterfacial tensions of alkane-water interfaces. The work is extended here to includethe interface between decan- l-ol and aqueous electrolytes.In many systems where the behaviour of liquid-liquid interfaces plays animportant role (e.g., emulsions, micellar solutions, artificial lipid membranes inwater), it is quite common for electrolytes to be present.Although, unless a systemcontains ionic surfactant, inorganic salts may not have a drastic effect, it is clearlydesirable to have some knowledge of the magnitude of the effect, and, ideally, a nunderstanding of its origins.At a decanol-water interface it is likely that the polar groups of the alkanol areclosely packed.The influence of salts, present in the water, on such an interfacemight be relevant in, say, the investigation of changes in the critical micelle concentra-tions of non-ionic detergents when salts are added. On the other hand, the way inwhich isolated alkanol molecules (at say an alkane-water interface) respond to thepresence of salts might be different. We have, therefore, additionally studied theadsorption to the dodecane-electrolyte interface of (a) butan-1-01 from aqueousNaCl, and (b) dodecan- l-ol from n-dodecane.EXPERIMENTALInterfacial tensions were determined at 20°C using the drop-volume technique aspreviously described,2 and were reproducible to better than 0.05 mN m-'.The interfacial tension between decanol and water was independent of the length oftime the two phases were left in contact. Also, in experiments carried out with NaCl andKI it was found that no detectable amounts of salt (using standard titration techniques)were present in decanol which had been in contact with aqueous electrolyte for 24 h ormore. Further, there was no indication of transfer of any of the salts to the decaiiol duringthe interfacial tension measurements.In studying the adsorption of butanol from aqueous NaCl, account had to be taken ofthe distribution of the alcohol between the aqueous phase and dodecane.Prior to thedetermination of each interfacial tension, approximately 10 cm3 alkane and 100 cm3 aqueous8R. AVEYARD, S. M. SALEEM AND R. HESELDBN 85butanol were placed in a stoppered flask and shaken mechanically for two days in a waterthermostat maintained at 20fO.l"C. Since the distribution ratio for butanol is about 9(using molarity units) in favour of ~ a t e r , ~ the aqueous phase concentration was reduced byabout 1 % and an allowance was made for this. These necessary procedures rendered theresults a little less reproducible than those (see below) for the adsorption of dodecanol, andso very low butanol concentrations (i.e., less than about 3 x mol fraction), where inter-facial tension is a linear function of bulk concentration, were not studied.For the adsorption of dodecanol from solution in dodecane, no pre-equilibration betweenthe phases was necessary, and only bulk phase mole fractions of dodecanol up to about0 .7 ~ were used.Dodecane (Newton-Maine, purity > 99.5 %) was passed through chromatographicalumina prior to use. The decanol and dodecanol (Koch-Light, puriss), and butanol(Fluka, pztriss) had purities, as estimated by g.1.c. in this laboratory, of approximately99.3, 99.8 and 99.8 %, respectively. Water, taken from a laboratory still, was twicedistilled using all Pyrex glass apparatus. The samples of salt were the same as thosedescribed in ref. (1). In addition, BaCI2 (AnalaR grade) was used and was, like the othersalts, heated to 450°C for > 12 h in a silica crucible to remove organic impurities.RESULTSInterfacial tensions, yap, for decanol in contact with aqueous solutions of LiCl,NaCI, KC1, KBr, KI and BaCl, were determined at 20"C, and are given in table 1.All the salts except KI caused an increase in yap.TABLE 1 .-INTERFACIAL TENSIONS, yap, OF DECAN- 1 -OL AGAINST AQUEOUS ELECTROLYTESAT 20°Ceiectrolyte m,/mol kg-1 yaB/rnN m-1 electrolyte rn3lrnol kg-1 @/mN m-1NaCl 00.1230.2010.3430.5050.6650.81 30.9251.020KCl 0.2140.3950.6120.8381.146LiCl 0.1020.2040.3860.5970.7798.839.008.999.179.229.379.429.579.669.069.189.389.539.738.918.979.089.209.23KBr 0.2190.4190.6030.8421.025KI 0.2010.41 10.6220.8331.105BaClz 0.2010.4050.61 10.8210.9698.969.039.089.209.258.718.578 -458.368.299.089.259.479.669.83Results for the adsorption of dodecanol from dodecane to the interfaces withvarious electrolytes at 20°C are presented in table 2, as surface pressures, n, (thelowering of interfacial tension caused by adsorption) as a function of the mole fractionx of dodecanol in dodecane.For a given salt concentration n: and x are linearlyrelated. The surface pressures for butanol adsorbed from aqueous NaCl to theinterface with dodecane are listed in table 3. The mole fraction of butanol is definedas n(butanol)/[n(butanol) + n(water) + 2n(NaCI)], where the n are numbers of molesTABLE 2.-sURFACE PRESSURES, Z, FOR DODECAN-1-OL ADSORBED FROM DODECANE TO DODECANE-AQUEOUS1 0 4 ~ n/mN m-1dodecanol waterin dodecane0.082 0.140.105 0.170.207 0.350.212 0.310.302 0.450.364 0.540.425 0.710.480 0.700.683 1.010.688 1.04NaCl KC1 KBr K1m=1.020 m=2.082 m=1.031 m=2.129 m=1.038 m=2.157 m=1.050 m=2.2120.19 0.26 0.21 0.21 0.18 0.11 0.22 0.140.35 0.53 0.37 0.41 0.35 0.35 0.44 0.330.54 0.65 0.58 0.59 0.55 0.59 0.61 0.540.72 0.80 0.78 0.77 0.71 0.82 0.78 0.861.02 1.11 1.0s 1.09 1.03 1.03 1.11 1.0R .AVEYARD, S . M. SALEEM A N D R. HESELDBN 87TABLE 3.-sURFACE PRESSURES, IT, AND AQUEOUS PHASE MOLE FRACTIONy X , OF BUTAN-1-OLFOR ADSORPTION AT THE AQUEOUS NaC1-DODECANE INTERFACE AT 20°C1 0 4 ~1.031.401.892.643.323.794.324.625.00concentration of salt, m3/mol kg-10.78 ___--- 0 0.27 0.56nlmNm-1 10% n/mNm-’ lOlx n/rnNm-1 lO4x njmNm-12.45 0.33 1.21 0.29 1.15 0.28 1.623.10 0.55 1.84 0.57 1.83 0.60 2.653.81 0.85 2.47 0.89 2.64 0.88 3.255.10 1.11 2.W 1.15 3.20 1.13 3.856.02 1.41 3.59 1.83 4.62 1.83 5.316.57 2.04 4.73 2.11 5.09 2.12 5.867.21 2.29 5.42 2.49 5.78 2.40 6.417.57 2.84 5.95 2.98 6.68 3.01 7.547.911.14 -1 0 4 ~ n/mN m-10.27 1.820.52 2.430.79 3.051.05 3.941.39 4.712.21 6.512.51 7.162.76 7.60DISCUSSIONWe include data derived from previous results for electrolyte-air and electrolyte-dodecane interfaces,l as well as the present results for the electrolyte-decanolinterface.WORK OF ADHESIONOne possible way of expressing the effects of salts on material at interfaces is toexamine the work of adhesion, Wi@, between the organic phase, a, and the aqueouselectrolyte, phase p ; Wi@ is given bywhere y“ and y p are the surface tensions of a and p respectively.* In all the presentcases Wip varies in an apparently linear fashion with salt molality, m3, and can berepresented bywip = ya+yp-yae (1)where B is a constant and (Wip), is the work of adhesion of decanol or dodecane,with water.In the sense that B values, listed in table 4, are positive, the salts investi-gated can be said to “salt-in” both decanol and dodecane at the interface withTABLE 4.-vALUES OF B [EQN (2)] FOR THE WORK OF ADHESION OF DECAN-1-OL AND OFDODECANE WITH AQUEOUS ELECTROLYTES AT 20°CsaltLiClNaClKClKBrKIBaClzNa2S04BlmJ m-2 kg mol-1decanol dodecane1 .oo 0.020.94 0.290.82 0.220.96 0.501.74 1.311.93 -- 0.26* For the decanol-water systems, values of ya and yp for the non-mutually saturated liquids wereused.Values of fp for the saturated systems were not significantly different from (although morereproducible than) those obtained on initial contact of the two phases ; values quoted are for theequilibrium systems88 DESORPTION OF ELECTROLYTESaqueous electrolyte, the effect for a given electrolyte being greater for decanol thanfor dodecane. The changes are small, however. Suppose the area occupied by adecanol molecule at the interface is 0.20 nm2 (see later). Then, expressed per moleof decanol, the increase in Wib (for the decanol-electrolyte interface) caused by1 mol kg-1 NaCl is only 113 J mol-l, and by 1 mol kg-' BaClz is 230 J mol-'.The thermodynamic significance of Wif, and of the change, AWiB, caused by theaddition of salts to pure water, is not particularly simple, however. In general.y ina system of i components is given byY = A0-Z ripi ( 3 )iwhere A, is the specific excess interfacial Helmholtz free energy, Ti are surfaceexcesses, and pi are chemical potentials. It follows from eqn (1) and (3) that fortwo immiscible liquids, 1 (the sole component of phase a) and 2 (the solvent in phase p,which also contains the salt, component 3), Wifi is given byThe change, AWZp, is thuswhere it has been assumed that T;p is unchanged by the presence of salt, and I-';@ andI-'[ have been expressed relative to planes such that, respectively, I'gfi and are zero.Even this simplified expression for AWip contains a term for changes (denoted A)in A , caused by addition of salt, as well as a term in the difference in desorption ofsalt at the ap and interfaces.In a previous paper the contribution of the term inA, to AWip for the dodecane-electrolyte interface was estimated, crudely, by theuse of a theory of van der Waals forces.' In order to construct a simple physicalpicture of the nature of salt effects at interfaces, however, it is probably moreilluminating to consider values of F3 at the various types of interface.W f = A," + At - A",B + (qp- I';)pl + (I';p - r g ) p 2 + (r';B - r3p)p3.A W;P = A(AB, - A",B) + (ry - r , ~ ) ~ ,(4)(9ADSORPTION OF ELECTROLYTESFor the systems of present interest, in which liquids 1 and 2 are effectively im-miscible, and where the salt is soluble only in the water (liquid 2), it is readily shownthat the surface excess of salt, Tiz), relative to the plane where the surface excess ofwater, T,, is zero, is given byIn eqn ( 6 ) f k is the mean ionic molal activity coefficient of the electrolyte and v isthe number of moles of ions per mole of salt.As Ralston and Healy have pointedthe possibility exists that the planes corresponding to r2 = 0 may differ fordifferent concentrations of an electrolyte, and for different electrolytes. Thecomparisons of behaviour which follow are only as valid as is the assumption of aconstant I-', = 0 plane.In all cases y was observed to be a linear function of m3 f*, within experimentalerror.Activity coefficients were taken from the compilation in ref. (5) and are for2 5 T , but (in the present context) the coefficients for 20°C are very unlikely to besignificantly different. Values of for 1 in01 kg-' electrolytes are given in table 5.Ions are repelled by electrical image forces from an interface between aqueouselectrolyte and a second phase of low relative permittivity, E , . ~ , Further, it isbelieved that ions at the air-electrolyte interface can be attracted or repelled byforces associated with the hydration of ions and of the interface. Although E , fordecanol is 8 and that for dodecane -2, electrostatic theory [e.g., that of BellemansTi2) = - m,f~(dy/dm,f*)/vRT.(6R. AVEYARD, S . M . SALEEM AND R. HESELDEN 89referred to in more detail in ref. (l)] predicts only very minor differences in the eleva-t ion of interfacial tension for the systems of present interest. We therefore explorethe possibility that the trends in our results, as seen in table 5, are a consequence ofhydration effects.TABLE 5.-vALUES OF rp’ FOR 1 mOl kg-’ ELECTROLYTES AT 20°C ainterface LiCl NaCl KCI KBr KI BaCl2air /electrolyte 3.36 3.53 3.36 2.86 2.54 4.09C ,,H,,Jelectrolyte 3.02 2.95 2.91 1.80 -0.15 2.58CloH210M/electrolyte 1.23 1.57 1.63 0.84 -1.05 1.41a Values listed are of - lO’r~~/rnol m-’.In fig. 1, Ti2) for 1 mol kg-’ electrolytes is plotted against & I I , the entropychange which accompanies the formation of the outer hydration cosphere 9 9 lo ofthe appropriate ions.The desorption decreases, at all three interface types, as the‘‘ structure-breaking ” propensity of the anions increases (i.e., as Sx,Ir becomes morepositive); the desorption is much less dependent on the state of hydration of thecations. A similar plot is obtained if the free energy of ion hydration is substitutedfor Sx,Ir and the observed effects could be associated with the properties of eitherthe inner or the outer cosphere, or of course both. For a given salt the desorptionis least for the decanol-electrolyte interface and greatest for the air-electrolyteinterface.4 r-N IEI- 4 0 0 4 0 6 0SX,II/J mol-’ K-’FIG.1 .-Surface excess of salts in I rnolal solution as a function Of Sx.11. Curve 1, air-electrolyte ;curve 2, dodecane-electrolyte ; curve 3, decanol4ectrolyte. Ti’’ for LiCI, NaCl, and KCl (0) areplotted against SXJI for the cations ; rS2’ for KCI, KBr and KI (0) are plotted against S ~ , I L forthe anions.Less is understood about the structure of water at interfaces than about thehydration of ions. It is probable that the surface monolayer (at least) of water at theair-water surface is oriented such that on average the oxygen atoms point towardsthe vapour phase,’. “-14 although the extent of this orientation is not clear. None-theiess, the specific excess surface entropy, -dy/dT, for water is positive (0.26 mJm-2 K-I) suggesting that the structure over the whole interface is less extensive tha90 DESORPTION OF ELECTROLYTESin bulk.Good l4 has noted that the surface entropy of water is considerably morenegative than for many “normal” liquids, but this does not of course mean, asappears to be implied by Horne,15 that the surface structure is more extensive thanthat in bulk water.Both the enthalpy, AhiD, and the entropy, As:’, of adhesion for the dodecane-water interface and for the decanol-water interface are positive (table 6). If themajor contribution to Ahib and AsgP arises from changes associated with the structureof interfacial water, it may be inferred from the sign of the adhesion parameters thatwater at the air-water interface is less structured than that at either of the liquid-liquid interfaces.From the magnitude of the parameters it would appear that the“ randomness” of the interfacial water decreases in the orderair-wa ter > dodecane-water > decanol-water.TABLE 6.-VALUES OF (w~b)o, Ah:’, AND AS;’ FOR DECAN-1-OL-WATER AND DODECANE-WATERINTERFACES AT 20°Cdodecane-water decanol-water( Wib)o/mJ m-2 45.2 92.4AhIP ImJ m-2 92.4 182.6As;p/mJ m-2 K-I 0.16 0.31Values taken from ref. (16).The assertion that the enthalpy and entropy changes largely reflect changes in waterstructure has no claim to rigour, but it is supported to some extent by the observationthat the ratios of the free energies, of the enthalpies, and of the entropies, for thedodecane-water and decanol-water interfaces are very nearly equal (ie., 0.49, 0.51and 0.52 respectively).This implies that a similar process, taking place to differentextents, is responsible for the thermodynamic parameters of adhesion for the twointerface types; this could be the disruption of the structure of interfacial water.If so, it is noteworthy that, although decanol is capable of H-bond formation withwater and dodecane is not, the kind of structure involved is apparently similar(thermodynamically) in both cases.that, for the three interfaces of present interest, the halideions are desorbed less strongly than the metal ions. Thus, if the above argumentsabout interfacial water are accepted, it is concluded that the strongest attractionoccurs between the least strongly hydrated (i.e. greatest structure-breaking) anionsand the interfaces with the most strongly structured water.has, quiterightly, observed that for the air-water interface, the structure-breaking anions arein a region of relatively disordered water. We do not believe, however, that theattraction is a result of this disorder per se, but rather of the average orientation ofthe water molecules in such structuring as does exist. Indeed, it is possible that thewater at the decanol-water interface is more structured than bulk water since theentropy of interface formation (-dy“B/dT) is negative in this case. If the same kindof surface orientation as exists at the air-water surface is enhanced by the presenceof the organic phase, the anion-interface attraction would be correspondinglyenhanced, and the desorption reduced.Accepting that the hydrogen atoms of water molecules at the air-water surfacepoint on average into the liquid, it is understandable in a general way that anionsshould be attracted to the surface, and that cations should be repelled.From ourresults it appears that water molecules at the alkane-water and decanol-waterinterfaces are similarly oriented, but to a greater degree. What is less clear is why,for a given type of interface, the anion-interface attraction is greater the weaker theThere is evidence ‘*RandleR. AVEYARD, S . M. SALEEM A N D R . HESELDEN 91ion hydration. More specifically, it is not known if the effects arise from the natureof the outer or the inner cosphere, or from both. However, it is probable that thelayer of water molecules, separating the anions at the interface from the non-aqueousphase, is only one or possibly two molecules thick on average.l- It may be,therefore, that the oriented surface layer of water is directly adjacent to anions andhence acts as a partial, primary hydration sheath.Under these circumstances anion with a smaller primary hydration number in bulk (e.g., I-)2o would be attractedmore to the interface than an ion (e.g., Cl-) with a greater hydration number.Further speculation on the basis of the present results would be premature.Finally in this section we refer to the work of Ralston and Healy who foundthat the desorption of KCl (from a 3 mol dm-3 solution at 21°C) at the air+lectrolyteinterface is reduced by a factor of 2 when a close-packed monolayer of octadecan-1-01(area per molecule = 0.20 nm2) is spread at the interface.These findings areentirely consistent with our results for the desorption, from 1 mol kg-' KC1, at theair-electrolyte and decanol-electrolyte interfaces (table 5). The desorption in thepresence of the spread monolayer was found to be very dependent on the surfaceconcentration of the alkanol, which supports our earlier suggestion that the area perdecanol molecule at the interface is about 0.20 nm2.ADSORPTION OF ALKANOLSThe influence of salts on d h t e adsorbed films of alkanols has been investigatedin terms of the standard free energy of adsorption, AapO, of dodecan-1-01 from dilutesolution in dodecane to the interface with aqueous electrolyte.The free energy isgiven by 21A a p = - RT In (nlx) (7)where (n/x) is the slope of the linear plot of .n against x, obtained by least squarestreatment of the data in table 2. The standard state for the surface is n; = 1 mN m-',and for the bulk the state where the product of mole fraction and activity coefficientof the solute is unity. Since the salts used are insoluble in dodecane, the effect ofsalt on Aape arises entirely from changes in the standard chemical potential, p*sb,of alkanol at the interface. Further, we may reasonably assume that any observedeffects are predominantly associated with the OH group of the alkanol rather thanthe hydrocarbon chain.TABLE 7.-sTANDARD FREE ENERGIES OF ADSORPTION, A@*, OF DODECAN-1-OL FROM DODECANETO THE INflERFACE WITH AQUEOUS ELECTROLYTES, MOLALITY m3, AT 20°Csalt NaCl NaCl KCI KCI KBr KBr KI KIrn3/mol kg-l 1.020 2.082 1.031 2 .1 2 9 1.038 2.157 1.050 2.212-A@/kJmol-' 23.35 23.4, 23.45 23-50 23.39 23.56 23.52 23.61salt BaC12 BaClz CaC12 CaClzrn3/mol kg-l 1.033 1.578 1.030 2 . 1 1 4-Aa/.L*/kJ mol-' 23.66 23.91 23.68 24.24-Aa@ for adsorption at the dodecane-water interface is 23.42 kJ mol-I.It is difficult to assign precise errors to the values of Asp* given in table 7 ; thepresent value for adsorption to the alkane-water interface is about 1 % more negativethan a previously determined value,22 obtained using a different sample of dodecanol.We believe that for a given sample however, Aape values are reliable to better than0.1 kJ mol-I in a relative sense.On this basis it is seen that the 1 rnol kg-l (approx.92 DESORPTION OF ELECTROLYTES1,l electrolytes investigated have very little effect on Aape. At -2 mol kg-l theeffects are more discernable ; KCI, KBr and KI cause to be more negative(corresponding to “ salting-in ” of dodecanol at the interface) by about, respectively,0.08,0.14 and 0.19 kJ mol-l. We estimate from the results of Wilcox and S ~ h r i e r , ~ ~that the salting-in of the OH group in alkan-1-01s in bulk aqueous solution (at 25°C)by 2molkg-l NaC1, NaBr and NaI corresponds to changes in free energy of,respectively, about -0.34, -0.30 and -0.25 kJ mol-l. It appears then, that forelectrolytes that are strongly desorbed (e.g., alkali metal chlorides), the salt effect ORalcohol at the interface is less than that on the OH group in bulk.As the saltsbecome progressively less desorbed, the influence of salt on the solute at the interfacebecomes more pronounced and approaches the bulk behaviour. The magnitudeof the free energy changes is small, however, and the conclusions drawn are tentative.The effects caused by BaC1, and CaC1, are more substantial; in both cases thealkanol is salted-in at the interface. Any possible effects that may result fromchanges in the (very low) solubility of dodecanol in water as a result of salt additionhave been ignored in the above discussion.The change in A,p* caused by NaC1, for the adsorption of butan-1-01 fromaqueous electrolyte to the dodecane-water interface is substantial. In this case,however, the salt also influences the chemical potential, p*J, of butanol in bulksolution and we now proceed to show that, in fact, virtually all of the change inAsp* arises from this change in p*J.Since the adsorption was not studied in theconcentration range where n and x are linearly related, Aap* could not be calculatedfrom eqn (7) and was determined as follows. It is known 2 2 that adsorbed mono-layers of alkan-1-01s at alkane-water interfaces obey the Volmer equationin which a is the area per molecule of adsorbate at the interface and a, is the molecularco-area. The present data for butanol at the alkane-water interface have beenanalysed as described in ref. (22) and found to conform to eqn (8) with a.= 8.27 nm2molecule-l. We have assumed the same is true for films at the alkane-electroryteinterface and calculated a from n (table 3) using eqn (8). It has already been shownthat the effect of 1 molal NaCl on pego for dodecanol (table 7) is very small. Ifsuch changes as do occur involve mainly the OH group of the alkanol, we expect thechange in peso for butanol to be very small also. If this reasoning is sound, thechange in Aape should be almost equal to the change in p.9” caused by salt, whichquantity is known independently.n(a-ao) = kT (per molecule) (8)The adsorption isotherm corresponding to eqn (8) is l9a0 A a p v x=- a0 exp - exp (- - 1).a-a, a-a, RT (9)The adsorption free energy Aapesv is for a surface standard state of (ideal) halfcoverage (k, aoJa = 0.5), and is related to Asp* [eqn (7)] by 21AapOsV = Asp* - RT[ln (ao/kT) - 11.(10)TABLE 8 .-STANDARD FREE ENERGIES OF ADSORPTIONy Aape, OF BUTAN-1-OL FROM AQUEOUSNaCI, MOLALITY m3, TO THE INTERFACE WITH DODECANE AT 20°Cm3/mol kg-l 0 0.27 0.56 0.78 1.14-A,iCLe/kJ mol-l 24.88 25.23 25.43 25.80 25.98Sample plots according to eqn (9) are depicted in fig. 2 and the free energy ofadsorption, values of which are given in table 8, is a linear function of sodiuR . AVEYARD, S. M. SALEEM AND R . HESELDEN 93chloride concentration with dA,,@/dm, = -0.98 If: 0.10 (R.M.S. error) kJ inol-'/mol kg-l. On the above arguments dA,pe/dm, should be nearly equal to pOgl(butanol in water)--**' (butanol in 1 mol kg-l electrolyte), the mole fraction ofbutanol being the same in water and in electrolyte.The value for this difference inchemical potential obtained from distribution experiment^,^^ has been found to be- 1.07 kJ mol-'.FIG. 2.-Isotherms for adsorption of butanol from aqueous NaCl to the interface with dodecane at20°C. 0 , H20 ; m, 0.56 mol kg-' NaCl ; 0, 1.14 mol kg-l NaCI.SUMMARYThe salts studied cause an increase in the work of adhesion between water andboth dodecane and decanol, and in this sense can be said to salt-in these organicliquids at the electrolyte-liquid interface.It is suggested that trends observed in the desorption of salts are a result of inter-actions between anions and the hydration layers at the interfaces. The strongestattractive forces exist between interfaces with the most structured water, and thegreatest structure-breaking (i.e., least strongly hydrated) anions.The salts studied have only a small effect on the chemical potential of aikanolsadsorbed at the alkane-water interface. This is particularly true of salts which arestrongly desorbed. For less strongly desorbed salts, the effect on p e y b approachesthat which the salt has on the OH group in bulk aqueous solution.The effect of NaCl on the free energy of adsorption of butanol to the alkane-electrolyte interface is substantial and arises almost entirely as a result of the salting-out of alcohol from aqueous solution.The authors thank the S.R.C. for the provision of a postdoctoral maintenancegrant (Colloid Science) for S.M. S . , and Unilever Research, Port Sunlight, for astudentship for R. H.R. Aveyard and S. M. Saleem, J.C.S. Faraday I, 1976, 72, 1609.R. Aveyard and D. A. Haydon, Trans. Faraday Soc., 1965, 61,2255.R. Aveyard and R. W. Mitchell, Tram. Faraday SOC., 1969, 65, 264594 DESORPTION OF ELECTROLYTESJ. Ralston and T. W. Healy, J. Colloid Interface Sci., 1973, 42, 629.R. A. Robinson and R. H. Stokes, Electrolyte Solutions (Butterworth, London, 1955).J. E. B. Randles, Disc. Faraday SOC., 1957, 24, 194.B. E. Conway, J. Electroanalyt. Chem., 1975, 65,491.A. Bellemans, Physica, 1964, 30, 924.H. S. Frank and W. Y. Wen, Disc. Faraday SOC., 1957, 24, 133.London, 1973), vol. 3, p. 1.and Tobias (Interscience, N.Y., 1963), vol. 3, p. 1.lo H. L. Friedman and C. V. Krishnan, Water, A Comprehensive Treatise, ed. F. Franks Wlenum,l1 J. E. B. Randles, Advances in Electrochemistry and Electrochemical Engineering, ed. Delaheyl 2 H. J. M. Nedermeijer-Denessen and C. L. De Ligny, Electroanalyt. Chem., 1974, 57, 265.l 3 N. H. Fletcher, Phil. Mag., 1962, 7, 255.14R. J. Good, J. Phys. Chem., 1957, 61, 810.R. A. Horne in Water and Water Pollution Handbook, ed. Caccio (Dekker, N.Y., 1972),chap. 17 : see in particular fig. 18.l6 R. Aveyard, B. J. Briscoe and J. Chapman, J.C.S. Faraday I, 1972, 68, 10.l7 R. Lumry and S . Rajender, Biopolyrners, 1970, 9, 1125.l8 D. A. Haydon, Biochem. Biophys. Acta, 1961, 50,457.l9 K. Johansson and J. C. Eriksson, J. Colloid Interface Sci., 1974, 49, 469.*O J. O'M. Bockris and A. K. N. Reddy, Modern Electrochemistry (Plenum, N.Y., 1970), vol. 1,21 R. Aveyard and B. J. Briscoe, Trans. Faraday Soc., 1970, 66,2911.22 R. Aveyard and B. J. Briscoe, J.C.S. Faraday I, 1972, 68, 478.23 F. L. Wilcox and E. E. Schrier, J. Phys. Chem., 1971,75, 3757.24 R. Aveyard and R. Heselden, J.C.S. Faraday I, 1975, 71,312.p. 131.(PAPER 6/1238