Solute-Solvent Interactions on the Surface of Reverse Phases Interactive Characteristics of some Short-chain Aliphatic Moderators having Different Functional Groups BY R. P. W. SCOTT* Chemical Research Department Hoffman-La Roche Inc. Nutley New Jersey 071 10 U.S.A. AND C. F. SIMP SON^ School of Molecular Sciences University of Sussex Brighton BNl 9HR Received 27th August 1980 The desorption-adsorption coefficients and distribution coefficients of a series of aliphatic alcohols carboxylic acids and aldehydes between water and ODs2 reverse phase have been determined using a chromatographic procedure. The moderator concentrations over which the measurements were made were kept at sufficiently low levels so that moderator/moderator interactions in the mobile phase and moderator interactions with adsorbed moderator were kept minimal.It is shown that the desorption-adsorption coefficients decrease and the distribution coefficients between water and the reverse phase increase exponentially with the carbon number of the moderator confirming the system can be described by the Martin equation. As a result substances having chains of five carbon atoms or more could produce complete coverage of the surface when present at a level of only a few percent in the mobile phase. Thus the surface of the reverse phase can be strongly modi- fied without significantly affecting the solvent characteristics of the mobile phase. It was shown that dispersive interactions between the reverse phase and the hydrocarbon chain of the moderator were the same for each moderator and from the slope of the log K against carbon-number curves it was also shown that the logarithmic increment for each methylene group in the alkyl chain was the same for all moderators and independent of the functional group.The difference in the distribution co-efficients between different moderators having alkyl chains of the same length depends solely on the dispersive interactions between the functional groups with the reverse phase and the functional group with any residual silanol groups that may be present. The method used can also determine the effec- tive chromatographic surface area of a given reverse-phase column. It is suggested that the technique could also be used to determine the effective surface area of different types of bonded phases give a measure of the residual polarity of the stationary phase and the magnitude of the interactive capacity of the hydrocarbon portion.Previous work examined the association of aliphatic ion-exchange materials such as the alkyl sulphonates with the surface of a reverse phase and demonstrated that they were adsorbed as a monolayer exhibiting Langmuir-type adsorption isotherms. In the same publication the adsorption of solvents as a monolayer on the surface of the bonded phase was also demonstrated. As ion-exchange interactions can be intro- duced on the surface of the reverse phase by adsorption of a suitable ion exchanger it would be reasonable to assume that other types of interactions could also be intro- duced by the adsorption of long-chain aliphatic alcohols acids or esters.For the system to be chromatographically effective the interacting agent would have to be held sufficiently strongly to the surface of the bonded phase to provide stability and the extent to which it was held would depend upon the chain length of the aliphatic portion. It was therefore of interest to determine the desorption-adsorption co-* Present address Perkin-Elmer Inc. Norwalk Connecticut U.S.A. t Present address Chelsea College University of London Manresa Road London SW3 6LX. 70 SOLUTE-SOLVENT INTERACTIONS ON THE SURFACE OF REVERSE PHASES efficients together with the distribution coefficients of a series of aliphatic moderators of different chain length with a given reverse phase.From such data the effect of chain length on the desorption-adsorption coefficients of the adsorbed solvents could be determined and the concentration required for the solvent completely to cover the surface could be calculated. It is likely that the interactions of the alkyl chain of the moderator with the sta- tionary phase would be dispersive in nature and it would also be likely that this inter- action would be independent of the functional group. The contribution of the func- tional group however to the distribution coefficient and thus retention would vary from functional group to functional group. For this reason the relative dispersive interactions of the hydrocarbon chain of a series of alkyl modifiers of different chain length having different functional groups was investigated.THEORETICAL The relationship between the desorption-adsorption coefficient of a substance which exhibits a Langmuir-type adsorption isotherm with the corrected retention volume of that substance has already been discussed.2 However an alternative algebraic procedure needs to be developed to permit other factors of the chromato- graphic system to be included in the expression relating the corrected retention volume of the moderating solvent with the moderator concentration; for example the distri- bution coefficient of the moderator in pure water and the effective chromatographic surface area of a given column. f 1 squarecm -+ Consider 1 cm2 of surface carrying an adsorbed layer of moderator at a concentra- tion of C gcm-2 in contact with a liquid containing C g of moderator per cm3 of the solvent.Let the molecular weight of the moderator be A4 and the area covered by the moderator molecule when adsorbed on the surface be S. Area of surface exposed = 1 -CS NS where N is Avogadro’s number. (1) M The number of molecules N, leaving the surface will be proportional to the con- centration of adsorbed molecules. i.e. Nl =/?Cswhere /? is a constant at a given temperature. The number of molecules striking and adhering to the exposed surface will be proportional to the concentration of the moderator in the mobile phase. i.e. N2= u(l -C,NS/M)C where u is another constant for the same temperature. Under equilibrium conditions Nl = N2 or PC,= u(l -C,NS/M)C,.Thus aC -uC,NSC,/M = PC and Cs(P+ aNSC,/M) = uC,. or R. P. W. SCOTT AND C. F. SIMPSON where K is the net effective distribution coefficient of the moderator between the sta- tionary phase and the solution of the moderator in water where g = p/a = the desorption-adsorption coefficient of the moderator. Now Y’ = Kv where V’ is the corrected retention volume of the moderator when chromatographed on a reverse phase where the moderator is present at a concentration of C in water and v is the total chromatographically available surface area of the moderator in the column. Thus from eqn (3) Thus if 1/ Y’,is plotted against C and the intercept and slope are ,u and w respectively and v = g/p. (6) It is seen from eqn (5) that the desorption-adsorption coefficient can be calculated from the slope and intercept of a l/Y’ curve plotted against moderator concentration provided the surface area of the moderator molecule is known.Furthermore the effective chromatographic surface area of the column can be calculated from the ratio of the desorption-adsorption coefficient to the intercept w as shown in eqn (6). A value for the effective chromatographic surface area could be extremely useful in comparing two reverse-phase columns and this could be achieved by simply examining the same water/moderator system on each column and determining the slopes and the intercepts of the curves as indicated above. Rearranging eqn (2),the equation for the adsorption isotherm can be obtained Eqn (7) describes the well-known Langmuir isotherm curve which can be calculated from the experimentally determined desorption-adsorption coefficients molecular weight of the moderator the surface area of the moderator molecule and Avogadro’s number N.From this equation the concentration of moderator necessary to com-pletely cover a given fraction of the surface could also be calculated. It should also be noted from eqn (4) that at infinite dilution (i.e. C = 0) Thus as V’ = Kv the distribution coefficient of the moderator is given by K = l/g where K is now the distribution coefficient of the moderator between the stationary phase and pure water. 72 SOLUTE-SOLVENT INTERACTIONS ON THE SURFACE OF REVERSE PHASES EXPERIMENTAL The apparatus used has been previously described3 and consists of a Waters 6000 M pump a thermostatted column fitted with a Valco automatic injection valve having an injec- tion volume of 2mm3 a Waters refractometer detector model 401 arid an appropriate recor- der.Water was qirculated through the column jacket and through the detector jacket from a thermostat that was maintained at 25.75 0.05 "C. The supply from the pump was first passed through a 3 ft coil submerged in the thermostat prior to entering the Valco valve in order to preheat the mobile phase to the correct temperature. The outlet from the detector was connected directly to a 10or 25 cm3 burette the choice depending on the retention volume of the substances being measured. A flow rate of 1 cm3 min-I was used; a burette reading was taken on injection and a reading again taken at the peak maximum.The burette was read to within 0.02 cm3; duplicate runs were carried out and the average of the two readings taken providing the duplicate did not differ by more than 0.05 cm3. If a greater difference was noted further replicate runs were made. The reverse phase employed was ODs-2 manufactured by Whatman Inc. Clifton N.J. and was packed in a 25 cm long column 4.6 mm i.d. This type of bonded phase was the so called " bulk " or polymeric reverse phase which tends to give somewhat higher retention than the alternative " brush " type4 and as previously shown does not give anomolous retention characteristics at high water concentra- tions as the " brush " type phases do.5 The mobile phases used were made up by weighing a known mass of moderator into a 100 cm3 standard flask and diluting to 100 cm3 with water.Moderator concentrations were limited to a maximum of 4% wjv to ensure that changes in interactions in the mobile phase were maintained at a minimum and thus changes in retention volume resulted from changes in the surface characteristics of the stationary phase only. The pump was purged with each new mobile phase and 50 cm3 of the mobile phase pumped through the column to ensure equilibrium before measurements were taken The concentration of the sample was adjusted so that the minimum mass of moderator in- jected was used to provide a peak 1/3 full-scale deflection at the maximum detector sensitivity. Dead volumes were measured for each mobile phase composition by determining the reten- tion volume of sodium chloride contained as a 0.1%wiv solution in water.Three aliphatic series were investigated alcohols (methyl ethyl propyl and butyl) carboxylic acids (formic acetic and propionic) and aldehydes (acetaldehyde and propionalde- hyde). For the C1and C2members of each series concentrations of up to 2% of moderator in water were employed. However the C3 members of the series tended to be relatively insoluble leading to nioderator/moderator interactions in the mobile phase at much lower concentrations. For this reason the retentions of these respective moderators were measured over concentration ranges of up to a maximum of 0.5 or 1% w/v. The effective area of each moderator molecule was obtained by the method described by Amoore.6 Molecular models were made to scale and placed on photographic paper and exposed together with a square rep- resenting (on the same scale) 9 A.z The square and the outlines of the molecules were cut out and weighed.This procedure was carried out for the three possible axial positions of each molecule which represented the three extreme possibilities of the area that a molecule could present to the chromatographic surface of the bonded phase on contact. The results are shown in table 1. In the first three columns the surface areas that were measured in the three different positions are given and the fourth column is an average of the three. The average values were used for calculating the different chromatographic adsorption para- meters given in the appropriate equations.The results for methanol ethanol propanol and butanol are given in tables 2 3 4 and 5 those for formic acetic and propionic acid are given in tables 6,7 and 8 respectively whereas the results for acetaldehyde and propionaldehyde are given in tables 9 and 10. The results given in these tables for the alcohols acids and aldehydes shown as curves relating 1/V' against moderator concentrations are given in fig. 1,2and 3 respectively. It is seen that ex- cellent straight lines are obtained with correlation coefficients close to unity. At this point it would be of interest to comment on the change in the measured dead volume shown in tables 2-9 inclusive. The change in dead volume with moderator con- centration and between the different moderators is not novel.It has been discussed by R. P. W. SCOTT AND C. F. SIMPSON 73 TABLE 1.-sURFACE AREA OF MODERATOR MOLECULES IN A’ X -Y phase X -2 phase Y -2 phase position 1 of position 2 position 3 mean moderator minimum area formic acid 18.5 18.4 22.5 19.8 acetic acid 20.3 22.5 29.0 23.9 propionic acid 26.9 26.6 35.9 29.8 formaldehyde 12.0 16.0 16.3 14.8 acetaldehyde 18.0 22.0 22.1 20.7 propionaldehyde 20.9 28.6 29.2 26.2 methanol 12.8 18.5 15.1 15.5 ethanol 21.o 25.1 23.4 23.2 n-pro pano 26.7 31.6 30.9 29.7 n- butanol 27.1 39.4 38.4 35.0 TABLE2.-RETENTION-VOLUME DATA FOR METHANOL IN METHANOL+ WATER SOLUTIONS ~~~~~~ solvent retention dead adjusted retention composition/g cm-3 volume VR/Crn3 volume vm/cm3 volume VR/cm3 1/ V’R /~m-~ ~~~~~~ ~~ 0.0000 4.08 2.05 2.02 0.495 0.0025 3.95 2.05 1.90 0.526 0.0050 3.83 2.05 1.78 0.562 0.0100 3 -64 2.05 1.59 0.629 0.0150 3.47 2.05 1.42 0.704 0.0200 3.36 2.05 1.31 0.763 Index of determination for curve fit to the function y = A + BX 0.999 A = 0.495 B = 13.62 TABLE 3.-RETENTION-VOLUME DATA FOR ETHANOL IN ETHANOL f WATER SOLUTIONS adjusted solvent retention dead retention composition volume volume volume 1/ V’R /g CM-~ V&m3 V,,,/cm3 V&m3 /cM-~ O.oo00 8.4 2.05 6.35 0.158 0.0025 7.15 2.05 5.10 0.196 0.0050 6.25 2.05 4.20 0.238 0.0100 5.20 2.03 3.17 0.3 15 0.0150 4.545 2.03 2.52 0.398 0.0200 4.10 2.03 2.07 0.483 Index of determination for curve fit to the function y = A + SX,1.000 A = 0.156 B = 16.23 74 SOLUTE-SOLVENT INTERACTIONS ON THE SURFACE OF REVERSE PHASES TABLE 4.-RETENTION-VOLUME solvent retention /g cm-3composition V&m3 volume 0.00000 25.2 0.00336 12.92 0.00648 9.30 0.00964 7.35 0.01280 6.34 0.0200 5.OO DATA FOR n-PROPANOL IN n-PROPANOL adjusted dead retention volume volume Vm/cm3 V’&m3 2.1 23.10 2.06 10.86 f WATER SOLUTIONS l/VR /~m-~ 0.0433 0.0921 0.138 0.189 0.232 0.337 2.06 2.07 2.03 2.03 7.24 5.28 4.31 2.97 Index of determination for curve fit to the function y =1 A + BX,0.999 A = 0.0423 B = 14.80 TABLE5.-RETENTION-VOLUME DATA FOR n-BUTANOL IN n-BUTANOL -/- WATER SOLUTIONS adjusted solvent retention dead retention /g CM-~ composition VR/cm3 volume Vm/cm3 volume V’,/cm3 volume /~rn-~ 1/ V’R 0.0000 101.25 2.02 99.23 0.0101 0.0025 23.10 1.95 21.15 0.0473 0.0050 14.15 1.78 12.37 0.0808 0.0100 8.OO 1.76 6.24 0.160 0.0150 6.10 1.72 4.38 0.228 Index of determination for curve fit to the function y = A + BX 0.999 A = 0.0102 B = 14.63 TABLE 6.-RETENTION-VOLUME DATA FOR FORMIC ACID IN FORMIC ACID 4-WATER SOLUTIONS adjusted solvent retention dead retention composition /g cm-3 volume V&m3 volume Vm/cm3 volume V’RICM3 1/ V’R /~m-~ 0.00000 3.05 1.90 1.15 0.869 0.00229 3.24 2.15 1.09 0.917 0.00458 3.15 2.08 1.07 0.935 0.00916 3.05 2.08 0.97 1.031 0.01 37 3.OO 2.08 0.92 1.087 0.0200 2.96 2.08 0.88 1.136 Index of determination for the curve fit to the function y = A + SX 0.986 A = 0.877 B = 13.56 R.P. W. SCOTT AND C. F. SIMPSON TABLE 7.-RETENTION-VOLUME DATA FOR ACETIC ACID IN ACETIC ACID f WATER SOLUTIONS solvent retention dead adjusted retention /g CM-~ composition V&m3 volume Vm/cm3 volume V’R/Crn3 volume /cmd3 1/ V’R 0.0000 6.00 1.95 4.05 0.247 0.0025 5.45 1.95 3.50 0.286 0.0050 4.98 1.95 3.03 0.330 0.0100 4.45 1.95 2.50 0.400 0.01 50 4.08 1.95 2.13 0.469 0.0200 3.80 1.95 1.85 0.541 ~~~~~ Index of determination for the curve fit to the function y = A + BX 0.999 A = 0.251 B = 14.60 TABLE DATA FOR PROPIONIC ACID IN PROPIONIC ACID + WATER 8.-RETENTION-VOLUME SOLUTIONS adjusted solvent retention dead retention composition volume volume volume 1/ V‘R /g cm-3 V&m3 Vm/cm3 v’R/cm3 /~rn-~ 0.0000 17.15 1.95 15.20 0.0658 0.0025 11.47 1.90 9.57 0.1045 0.0050 8.85 1.87 6.98 0.143 0.0075 7.50 1.90 5.60 0.179 0.0100 6.42 1.85 4.57 0.219 Index of determination for the curve fit to the function y = A 4-SX 1.000 A = 0.0656 B = 15.26 TABLE 9.-RETENTION-VOLUME DATA FOR ACETALDEHYDE IN ACETALDEHYDE + WATER SOLUTIONS adjusted solvent retention dead retention composition volume volume volume / V’R /g ~rn-~ V&m3 Vm/cm3 vycm3 /~m-~ 0.0025 5.72 1.97 3.75 0.267 0.0050 5.15 1.97 3.18 0.314 0.01 00 4.44 1.98 2.46 0.406 0.01 50 4.02 1.96 2.06 0.485 0.0200 3.75 1.95 1.80 0.555 Index of determination for the curve fit to the function y = A + BX 0.996 A = 0.232 B = 16.53 76 SOLUTE-SOLVENT INTERACTIONS ON THE SURFACE OF REVERSE PHASES Shoenmaker et aZ.,’ by Knox and Jurand’ and also by McCormick and Karger who have examined the phenomena in detail.The retention volume (VR) of a substance can be given by the following equation VR = Vi + K~vp+ &A where Vi is the interstitial volume V is the pore volume A is the surface area of the sta- tionary phase Kl is the distribution between the mobile phase and the stationary phase in the TABLE 1~.-RETENTION-VOLUME DATA FOR PROPIONALDEHYDE IN PROPIONALDEHYDE + WATER SOLUTIONS adjusted solvent retention dead retention /g cm-3 composition vR/cm3 volume V,,,/cm3 volume V’&m3 volume /~rn-~ 1/ V’R ~~~ 0.001 25 13.30 1.85 11.45 0.087 0.002 50 11.10 1.86 9.24 0.108 0.003 75 9.60 1.85 7.75 0.129 0.005 00 8.50 1.80 6.70 0.149 Index of determination for the curve fit to the function y = A + SX,1.0oO A = 0.0665 B = 16.56 pores and K2is the overall distribution coefficient between the surface and the mobile phase.If a solute is chosen such that K2 = 0 then VR = Vi + Klvp and if the liquid in the pores is the same as the mobile phase then Kl =r 1 and VR = Vi f vp = VO where Vo is the classically defined dead volume. In experiments given in this paper the adsorption of the moderator onto the surface which exists inside the pores can cause both the volume of the pore to change and also the solvent composition of the liquid in the pores so that Kl may not necessarily equal 1 at all times.Because a salt is used as a dead-volume marker however it is likely that K, if not unity is very close to unity. However the pore volume will change significantly and this change will be the greater where the moderator is strongly adsorbed and this is shown for propionic acid propionaldehyde and butanol in tables 5,s and 10. The change in dead volume will be the least for the least strongly adsorbed molecules for example methanol as shown in table 2 where the dead volume does not change at all and for acetaldehyde which also shows very little change. DISCUSSION OF RESULTS The linear relationship between 1/ V’ (the reciprocal of the corrected retention volume) and the moderator concentration shown in fig.1-3 confirm that the moderators are being adsorbed onto the surface of the reverse phase according to a Langmuir-type adsorption isotherm. The correlation coefficient of the curves relating mobile phase composition and the reciprocal of corrected retention volume together with the constants A and B which are the intercept and slope of the curve respectively are shown for all three series of moderators in table 11. The value of NS/M was calculated using the data in table 1 Avogadro’s number and the molecular weight of the alcohol. It is seen that although the mean molecular TABLE OF CHROMATOGRAPHIC DATA 11.-SUMMARY effective surface distribution intercept adsorption-desorption area of column coefficient NS/M IU slope coefficient 0 K /lo-' cm2 /~m-~ lg-' 1104 g Im2 cm-l formic acid 2.60 0.877 14.79 15.4 176 0.649 acetic acid 2.40 0.251 14.60 41.3 165 2.42 propionic acid 2.43 0.0656 15.26 10.4 I59 9.57 acetaldehyde 2.84 0.232 16.53 39.8 172 2.5 1 propionaldehyde 2.73 0.0665 16.56 10.96 165 9.122 methanol 2.90 0.495 13.62 105 21 3 0.949 ethanol 3.04 0.156 16.23 29.2 187 3.42 propanol 2.98 0.0423 14.80 8.52 201 11.7 n-butanol 2.95 0.0 102 14.63 2.05 202 48.6 78 SOLUTE-SOLVENT INTERACTIONS ON THE SURFACE OF REVERSE PHASES I 0 0.02 concentration of moderator 1g cm-3 FIG.1.-Plot of the reciprocal of the corrected retention volume of four aliphatic alcohols and their concentration in the mobile phase. (a) MeOH I.D. 0.999; slope 13.62; intercept 0.495.(b) EtOH I.D. 1.O00; slope 16.23; intercept 0.156. (c) PrOH I.D. 0.999; slope 14.80; intercept 0 0423. (d) BuOH I.D. 0.999; slope 14.63; intercept 0.0102. 0.8 n I E -2 0.6 9 b . H W 0.4 0.2 I I 0 05 1.0 15 2.0 concentration of moderator (acid) (% wlv) FIG.2.-Plot of l/V for three aliphatic acids against moderator concentration. (a)Formic acid I.D. 0.986; slope 14.79; intercept 0.877. (b)Acetic acid I.D. 0.996; slope 14.60; intercept 0.251. (c)Propionic acid I.D. 1.O00; slope 15.26; intercept 0.0656. R. P. W. SCOTT AND C. F. SIMPSON areas differ considerably between each moderator the value of NS/Mfor a given series is very similar for each member. This means that the mean surface area increases linearly with the molecular weight which could be expected; thus the surface area per gram for each respective moderator remains approximately constant.The de- sorption-adsorption coefficients were calculated using eqn (5). It is seen that the desorption-adsorption coefficients vary from a value of 105 x lo4 for methanol to 2.05 x lo4 for n-butanol for the alcohol series. Eqn (8) permits the calculation of the 0.6 1 0 6-5 1.0 I I .5 I 2.0 concentration of moderator (aldehyde) (% w/v) FIG.3.-Plot of 1/V for two aliphatic aldehydes against moderator concentration. (a)Acetaldehyde I.D. 0.996; slope 16.53; intercept 0.232. (b)Propionaldehyde I.D. 0.998; slope 14.98; intercept 0.017. distribution coefficient of the solvent at infinite dilution and is given by the reciprocal of "g " and shown in the last column of table 11.From the values of "g " given in table 11 and by use of eqn (7) the adsorption isotherms for the four alcohols were calculated and are shown in fig. 4. It is seen that for n-butanol and n-propanol over 95% of the surface is covered with the alcohol at a concentration of ca. 7% w/v. However at 7% w/v of ethanol only 86% of the surface is covered and with methanol at 7% w/v only ca. 67% is covered. It is fairly obvious that the adhesion of the alcohol layer to the hydrocarbon chain of the reverse phase increases rapidly with carbon chain length. In fig. 5 the distribution coefficient (K)of each member of each series is plotted on a logarithmic scale against its respective carbon number.It is seen that the curves for the homologous series of acids and alcohols give excellent straight lines having indices of determination close to unity confirming the compliance of the system with the Martin equation.l0 These curves are also in agreement with those obtained by Colin and Guiochon l1 and Berendsen.12 These authors however showed a linear relation- ships between log k' and carbon number which of course includes the surface area of the support and is thus not solely related to the interactions of the solvent with the two phases. The slopes and intercepts for the two linear curves relating log K against carbon number are shown in table 12. Included in table 12 is the slope of the line 80 SOLUTE-SOLVENT INTERACTIONS ON THE SURFACE OF REVERSE PHASES 41 complete coverage of surface by monolayer EtOH c , I 1 I I 0 0.02 0.04 0.06 0.08 0.1 concentrationof alcohol in mobile phaselg cm-3 FIG.4.-Adsorption isotherms for the C1 to C4alcohols on ODs2 reverse phase.; 3 carbon number FIG.5.-Plot of log K against carbon number for a homologous series of moderators having different functional groups. (a)Alcohols I.D. 0.999; slope 1.304; intercept 0.251 x (6)Aldehydes I.D. 1.0oO; slope 1.290; intercept 0.691 x (c) Acids I.D. 1.000; slope 1.345; intercept 0.643 x R. P. W. SCOTT AND C. F. SIMPSON joining the points for acetaldehyde and propionaldehyde together with the intercept found by the projection of the line to the log K axis.It is seen from table 12 that the slope of the log K against carbon number curves for all three series are numerically very similar clearly demonstrating that the contribu- tion to log K of each methylene group is independent of the terminal functional group of the series which again would be expected from a system that complied with the Martin equation. Thus the difference in retention volumes between an alcohol aldehyde or acid having the same number of carbon atoms in their aliphatic chain is TABLE 12.-sUMMARY OF DISTRIBUTION DATA slope intercept intercept of log K residual value for K moderators per carbon number due to functional groups alcohol 1.30 0.251 aldehydes 1.29 0.691 acids 1.35 0.643 mean 1.31 standard deviation 0.032 determined solely by the nature of the functional group.Algebraically the relation- ship between the distribution coefficient of an aliphatic substance between a reverse phase and water can be given by the following equation logK= A + 1.31 n (9) or K = Be's3'" when 1.31 is the mean slope from table 8 n is the number of carbon atoms in the side chain A is a constant depending on the functional group and A = log B. The relationship given in eqn (9) will be temperature dependent and also will vary with the nature of the stationary phase. If a specific type of stationary phase is em- ployed such as the bulk-modified materials used in this paper then the equation will hold for other stationary phases providing they have the same carbon content. If the carbon content changes this will also modify the constants in eqn (9).The functional group/stationary phase interactions will be largely dispersive in nature but due to the non-ideality of practical reverse phases polar interactions with any urireacted silanol groups will also be included. It is therefore not possible to relate the dispersive interactions of the functional group solely to the intercepts given in table 12 and consequently they are not related directly with the constant B in eqn (9). It is also interesting to note from table 11 that the surface area available to the different moderators although similar is not identical and it would appear that the alcohol was exposed to a somewhat greater area of surface than the acids and alde- hydes. Combining eqn (5) and (6) it is seen that V = NS[yM (10) and thus the calculations of the surface area p depend on the accuracy by which the surface area of the molecules S is determined and thus hinges on surface area measurement techniques suggested by Am~ore.~ It is likely that although Amoore's method gives approximately correct values for molecules errors are possible parti- cularly when different functional groups are involved.However as the calculation of the distribution coefficient depends on these area values and excellent correlation is 82 SOLUTE-SOLVENT INTERACTIONS ON THE SURFACE OF REVERSE PHASES demonstrated by the results given in fig. 5 with the Martin equation then the areas given in table 11 must also be valid as they are derived from the same basic equation.CONCLUSIONS It has been shown that adsorption isotherms for a number of aliphatic moderators on a reverse phase can be determined chromatographically. It has also been shown that the desorption-adsorption coefficient decreases exponentially with the carbon number of the moderator. Conversely the distribution coefficient of each molecule increases exponentially with the carbon number. It is clear that when using an ali-phatic modifier having a chain length of 4 or 5 carbon atoms the surface of a bonded phase could be completely covered with a monolayer and by choosing appropriately active groups the chromatographic characteristics of the surface could be changed. Providing sufficiently long carbon-chain species are used the modification can be accomplished at very low concentrations thus minimizing moderator interactions in the mobile phase.The technique described also permits the effective chromatographic surface area of a column to be determined and it has been shown that approximately 200 m2 of surface is chromatographically available from the ODs-2 column examined employing an alcohol as the surface probe. However ODs-2 reverse phase is poly- meric in nature and it may well be that a " brush " type reverse phase could have a significantly different effective surface area indeed the effective chromatographic surface area per gram could be one parameter by which a reverse phase could be characterized. It has been shown that dispersive interactions between the reverse phase and the hydrocarbon chains are similar for all the three series and independent of the func- tional groups.The distribution coefficient (K)of an aliphatic acid alcohol or alde- hyde between water and the reverse phase examined can be described by the following simple equation K = Be1.31" when B is a constant and is characteristic cf the functional group and nis the number of carbon atoms (methylene group) in the side chains. It follows that by using K for one member of a given aliphatic series the value of K for any other member of that series can be calculated. One of the authors (C.F.S.) would like to thank Whatman Inc. Clifton N.J. for support and Hoffman-La Roche Inc. for kindly providing laboratory facilities. Both authors wish to thank Ms.C. Lancaster and Mr. T. Garrison for help with the experimental work. R. P. W. Scott and P. Kucera J. Chronratogr. 1979 179 51. R. P. W. Scott and P. Kucera J. Chromatogr. 1978 149 93. R. P. W. Scott and P. Kucera J. Chromatogr. 1975 112 425. * I. Halasz and I. Sebastian Angew. Chem. Int. Ed. Engl. 1969 8 453. R. P. W. Scott and C. F. Simpson J. Chrornatogr. 1980 197 11. J. E. Amoore Ann. N.Y. Acacl. Sci. 1964 116 457. 'P. J. Shoemakers H. A. H. Billiet R. Tijssen and L. De Galan J. Clzronzatogr.,1978 149 519. J. H. Knox and J. Jurand J. Chromatogr. 1978 149 297. R. M. McCormick and B. L. Karger Anal. Chem. 1980 in press. lo A. J. P. Martin Biochem. SOC.Symp. 1949 3 4. H. Colin and G. Guiochon J. Chromatogr. Sci. 1980 18 54. l2 G. E.Berendsen Doctorate Thesis (Technische Hogeschool Delft University Press).