Influence of water in liquid—solid chromatographic systems on retention data

Journal of Chromafography, 316 (1984) 373-388

Elsevier Science Publishers B.V.. Amsterdam CHROMSYMP.

Printed in The Netherlands

470

INFLUENCE OF WATER IN LIQUID-SOLID TEMS ON RETENTION DATA II. FIXATION OF MONOLOCALIZED AND SILICA GELS

CHROMATOGRAPHIC

SYS-

SOLUTES WITH CLASS N SOLVENTS

C. SOUTEYRAND, M. THIBERT, M. CAUDE* and R. ROSSET Laboratoire de Chimie Analytique de l’lkole Supirieure de Physique et de Chimie de Paris, 10 rue Vauquelin, 75231 Paris Ceakx 05 (France)

SUMMARY

A new model for monolocalized solute retention is presented in which silicas are used as adsorbent and class N solvents as mobile phase (diisopropyl ether, 1,2dichloroethane, dichloromethane, chloroform, carbon tetrachloride and cyclohexane). It is based upon displacement of solvent molecules fixed simultaneously on free, bonded silanol groups and water molecules attached to free silanol groups by hydrogen bonding by a solute molecule. No distinction is made between solute molecules adsorbed on bonded silanol groups and those on free silanol groups deactivated by a monolayer of water. This model is valid for isoactivating solvents having a reduced water content lower than 0.13, corresponding to a water mole fraction on free silanol groups of 0.75. Application of the model to a large number of solutes shows good agreement between theory and experiment. The solute-solvent equilibrium constants on free silanol groups and bonded or free silanol groups deactivated by a water molecule are calculated. These values allow us to draw the following conclusions: bonded silanol groups have a lower energy than free silanol groups; the adsorption energy of a functional group depends on the proximity and spatial volume of other groups; the relative adsorption energies of class N solvents are in good agreement with the solvent strength parameter, so, defined by Snyder.

INTRODUCTION

Several solute retention models have already been proposed for liquid-solid chromatography (LSC). In 1968, a detailed model for LSC retention was proposed by Snyder’ including the role of the mobile phase in the separation. Three classes of solvents were considered: class N, the less polar solvents with LSC solvent strength values, so, in the range from 0.00 to about 0.40 (for alumina); class P, the more polar solvents with so 0.54.8; and class A B, the amphoteric polar solvents. A somewhat different retention model was presented a few years later by Soczewinskiz, but in 1974 Snyder3 demonstrated that the two models are essentially equivalent, and more 0021-9673/84/%03.00

0

1984 Elsevier Science Publishers B.V.

C. SOUTEYRAND

374

et al.

recently the Snyder-Soczewinski (S-S) model was described in full detail by Snyder and Poppe4. In 1973 a third model was presented by Scott and Kucera5. This (S-K) model differs almost diametrically from the S-S model. According to Scott6F7, normal chromatographic silica (thermally activated at 1%200°C) contains free silanol groups, and a molecule of water is adsorbed onto each silanol group. This monomolecular layer of adsorbed water, when not displaced by anhydrous methanol, constitutes the surface of the stationary phase on which solute or solvent molecules are adsorbed from the mobile phase. A critical comparison of the S-K model with the S-S model was discussed by Snyder and Poppe4. Recently, a comprehensive and detailed physicochemical model was presented by Snyder*. This model is based on either classical adsorbents, such as silica or alumina, or on polar bonded-phase packings, such as aminoalkyl, diol, etc .9. It considers displacement and localization phenomena’O as the primary contribution to retention. Few models take into account the solvent water content in a quantitative way. Therefore, we propose a simple retention model in LSC for isoactivating class N solvents and for monolocalized solute adsorption. Water is considered here as a polar solvent, so binary water-class N solvents will be studied. Isohydric or isoactivating solvents (solvents giving the same support activity) are defined as solvents having the same reduced water content’ l. THEORETICAL

duced silanol model. can be groups

In a previous paper” we have shown that, for class N solvents having a rewater content lower than 0.13, the water adsorption takes place at strong groups (or free silanol groups) and can be described by the simple Langmuir The silica gel surface can be visualized as in Fig. 1. The solute molecules, s, adsorbed on free silanol groups (a), on bonded silanol(0) or on free silanol deactivated by a monolayer of water (W).

ssss i wwwwssssss . . . . ..oooo

t

ssssts

wwwwssssss . . . . ..oooo

a

i

ssss wwwwssssss l ~~**.oooo

ssss \ wwwwssssss . . . . ..oooo

ssss WWWWSSSsSS

•*~*~*oooo

b

J

SSSS

i

wwwwssssss .~~**~0000

C

Fig. 1. Monolocalized solute retention mechanism for class N solvents. l = Free silanol group; 0 = bonded silanol group; w , = free silanol group covered by a water molecule; s = solute molecule; S = solvent molecule. a. b aid c: see text.

INFLUENCE

OF WATER

IN LSC SYSTEMS.

The monolocalized equilibria.

II.

375

retention mechanism can be represented by the following

(a) Interaction with free silanol groups (Fig. la) s, + s, z$ s. + s,

Here, subscript a refers to molecules adsorbed on free silanol groups, and m to molecules present in the mobile phase; S is a molecule of class N solvent. This equilibrium is characterized by the corresponding thermodynamic equilibrium constant

(1) where [a& and [a&,, are the activities of the ith species in the stationary (for free silanol groups) and the mobile phase, respectively. The activity of the ith species is defined as

[ails = Yi.abila [ail, = Yimbilm where yi,. and Yi,mare the activity coefficients of the ith species in the stationary (for free silanol groups) and mobile phase and ]Xi]aand (Xi]mare the corresponding mole fractions. (b) Interaction on bonded silanol groups (Fig. lb) Sm + s,, z$ s,, + s,

Here, subscript a’ refers to molecules adsorbed on bonded silanol groups: (2) (c) Interaction on free silanol groups deactivated by a monolayer of water (Fig. lc) s, + s, * SW+ s,

Here, subscript w refers to molecules adsorbed on free silanol groups deactivated by a molecule of water:

We have previously shown that the adsorption energy of water on bonded silanol groups is nearly the same as that for additional water on free silanol groups to which molecules of water are already attached by hydrogen bonding’ ‘; it seems reasonable to make the same assumption for the adsorption energy of the solute.

C. SOUTEYRAND

376

et al.

Therefore:

Thermodynamic studies of adsorption isotherms have shown12-16 that as the stationary phase becomes more ideal the adsorption phenomenon is more localized. So polar adsorbents, such as silica, may be assumed to be ideal Yi,a =

Yi,a’ =

Yi,w

=

1

where yi,a, an d yi,ware the activity coefficients of the ith species in the stationary phase for bonded silanol groups and for free silanol groups deactivated by a monolayer of water, respectively. In the mobile phase, the solute concentration is always very low, and taking the pure solvent S as the standard state, we can write: Ys,m =

1

= constant

Ys,m

Eqns. 1, 2 and 3 can then be transformed

into: (4)

K!j

=

I&&lm Ys,ml~slml~Sla’

K;h

=

Idvldn Ys,mlXslmlXSlw

(5)

With class N solvents the water concentration will be very small and consequently negligible; in the same way the amount of solute injected will be small and negligible. So we can write: bHzOla

+

kla

=

1

IXSla~ = 1 I&

= 1

I&

= 1

In that case, the following relationships can be derived from eqns. 4, 5 and 6, respectively K’f

=

Id3 Ys,ml&U -

(7) bH201a)

INFLUENCE

377

OF WATER IN LSC SYSTEMS. II.

KLj = ~

IXSIW

(9)

Ys,mlXslm

and we can write two equilibrium constants: KI = K:"y,,m

(10)

K2

(11)

=

K’Pysm

The s solute capacity factor, k’, can be defined as the ratio of the numbers of molecules in the stationary and mobile phase, respectively k’

=

%,a +

%a’ +

ns,w

(12)

n s.m

where ns+, ns,,,, and ns,w are the numbers of adsorbed solute molecules per gram of silica on free silanol groups, bonded silanol groups and free silanol groups deactivated by a monolayer of water, respectively; n s,mis the number of solute molecules contained in the dead volume of the column per gram of silica. We have also the following relationships

IX& = ns.,JN

(13)

lx,Id = ns,dl~

(14)

I-& = ns,dNm

(16)

where N, N’ are the number of free and bonded silanol groups, respectively, per gram of silica, nu+,a is the number of water molecules deactivating free silanol groups and N,,, is the number of solvent molecules contained in the dead volume of the column per gram of silica. Eqn. 12 can then be transformed into:

k’=$

m

N’ K1 + K2-% -

WI

-

K2h201a

1

(17)

This relationship can be written in the simplified form k’ = a - blxH20(a where a and b are positive constants (K, > K2).

(18)

C. SOUTEYRAND

378

et al.

Previously’ l the following relationship was demonstrated (19) where K is the water-solvent equilibrium constant for adsorption groups. So, combination of eqns. 17 and 19 gives

on free silanol

(20)

i.e., of the type:

k’

=

a+

hH201m

1 + ~IXH201m

(21)

Eqns. 18 and 21 lead to simple expressions for the dependence of the k’ values of monolocalized solutes on the water mole fraction in the stationary and mobile phase, respectively. We must emphasize that these two relationships are only valid for class N solvents having a reduced water content lower than 0.13. EXPERIMENTAL

Experiments were performed with a liquid chromatograph, Spectra Physics Model 8000 (Spectra Physics, Santa Clara, CA, U.S.A.). The UV detector was a Spectromonitor Model III (Laboratory Data Control, Riviera Beach, FL, U.S.A.). The temperature was kept at 25 f O.l”C with a constant-temperature water-bath. The solute retention times were obtained from two measurements and the results were averaged. Reproducibility of retention time was 1%. Stationary phases

Two porous silica gels (5 pm) were used: LiChrosorb Si 60 and Si 100 (E. Merck, Darmstadt, F.R.G.). The surface area and pore volume, respectively, are 482 mz g-l and 0.85 cm3 g-l for LiChrosorb Si 60 and 309 m* g-l and 1.15 cm3 g-l for LiChrosorb Si 100. The columns (250 x 4.8 mm I.D.) were packed with 2.00 g of LiChrosorb Si 60 and 1.7 g of LiChrosorb Si 100. The mobile phase flow-rate was 1 ml min-‘. Chemicals

The class N solvents and corresponding test solutes studied are listed in Table I. Dichloromethane, 1,Zdichloroethane, chloroform and carbon tetrachloride were of LiChrosorb grade and were purchased from Merck. Cyclohexane and diisopropyl ether were of Spectrosol grade and were obtained from SDS (Valdonne, Peypin, France). Test solutes were of analytical grade and were purchased from various producers.

INFLUENCE

379

OF WATER IN LSC SYSTEMS. II.

TABLE I CLASS N SOLVENTS AND CORRESPONDING

TEST SOLUTES

Solvent

Test solutes

Diisopropyl ether

Phenol, aniline, o-nitroaniline, p-nitroaniline, yl-2-propanol, 3-phenyl-1-propanol

1,2-Dichloroethane

Phenol, p-cresol, o-nitroaniline, I-phenyl-1-propanol, 3-phenyl-1-propanol

Dichloromethane

Same as with 1,2-dichloroethane

Chloroform

Phenol, p-cresol, @aphthol,

Carbon tetrachloride

Naphthalene, o-methylmethoxybenzene, p-methylmethoxybenzene, methoxybenzene, o-ethylnitrobenzene, o-methylnitrobenzene, p-methylnitrobenzene, nitrobenzene

Cyclohexane

Benzene, fluorobenzene, chlorobenzene, nonadecylbenzene,

I-phenyl-1-propanol,

l-phen-

I-phenyl-2-propanol,

aniline, o-nitroaniline, p-nitroaniline

naphthalene

Water content of the mobile phase The water contents of the solvents were measured coulometrically by the Karl Fischer titration method (Automate Bizot et Constant, Prolabo, France)“. Accuracy was about 5% for water contents of 2-10 ppm and 2% for those greater than 10 ppm. Dead-volume determination The dead volume was measured by using n-hexane as test solute. RESULTS AND DISCUSSION

The capacity factor values of test solutes measured at 25°C for waterdiisopropyl ether, -l,Zdichloroethane, dichloromethane, chloroform, carbon tetrachloride and cyclohexane as mobile phase are shown in Tables II, III, IV, V, VI and VII, respectively. The water mole fraction on free silica groups was calculated from the watersolvent equilibrium constant published previously’ l. According to eqn. 18 a linear decrease of capacity factor values is observed for all these solutes versus water mole fraction on free silica groups up to a maximum of 0.75. (In this range water adsorption takes place at free silanol groups only, see Figs. 24) Beyond this point, water adsorption takes place simultaneously at bonded silanol groups and free silanol groups having a water molecule already attached by hydrogen bonding. So, eqn. 18 is not valid, and a steep decrease in capacity factor values is observed (Fig. 2). A non-linear regression has been used to fit the experimental points with eqn. 21, giving the values sought. A short program for a Hewlett-Packard 9825 A computer has been used to obtain the a, /? and K values. As an example, water-chloroform equilibrium constant values calculated for six test solutes are shown in TableVIII. Aclose fit between the average value obtained from Table VIII (3400) and the value obtained from the experimental adsorption isotherm (3500)” is observed. These overall results support the validity of our theoretical model.

TABLE II

151

0.856

0.371

Water content in mobile phase (ppm)

IX”*Olrn lo3

Water mole fraction on free silanol groups, bHIOia

0.452

1.20

211

0.515

1.54

272

0.42 3.10 0.94 4.15 0.87 2.93 3.86

0.46 3.84 1.00 4.80 0.96 3.35 4.51

Phenol Aniline o-Nitroaniline p-Nitroaniline I-Phenyl- 1-propanol I-Phenyl-2-propanol 3-Phenyl-1-propanol

0.44 3.40 0.96 4.41 0.91 3.11 4.11

Capacity factor on LiChrosorb

Solute

0.616

2.33

411

0.40 2.69 0.90 3.73 0.80 2.62 3.43

Si 60

0.721

3.14

660

0.38 2.25 0.87 3.27 0.73 2.33 2.98

0.932

19.7

3480

0.32 1.02 0.71 2.00 0.49 1.32 1.57

0.215

0.397

70

0.34 2.11 0.655 3.42 0.66 2.31 3.12

0.342

0.754

133

0.315 2.41 0.62 3.04 0.60 2.05 2.14

Capacity factor

0.434

1.11

196

0.305 2.09 0.60 2.13 0.51 1.85 2.48

on LiChrosorb

0.558

1.83

323

0.28 1.82 0.57 2.43 0.51 I .68 2.20

Si 100

0.719

3.71

655

0.25 1.43 0.53 2.07 0.46 1.42 1.84

VARIATIONS OF CAPACITY FACTOR VALUES OF TEST SOLUTES VERSUS WATER CONTENT IN DIISOPROPYL ETHER

0.910

14.6

2580

0.20 0.69 0.43 1.21 0.31 0.84 1.00

41

0.225

35

0.193

0.307

lXHZOlmlo3

Water mole fraction on free silanol groups, IXH,Ol.

0.341

0.373

0.259

47

Si 60

0.406

0.297

54

0.975 2.09 2.67 2.87 6.74 13.5 16.5

Water content in mobile phase (ppm)

0.98 2.13 2.72 2.90 6.81 13.9 16.8

1.02 2.24 2.81 3.01 7.05 14.5 17.6

o-Nitroaniline p-Nitroaniline Phenol p-Cresol 1-Phenyl- 1-propanol I-Phenyl-2-propanol 3-Phenyl- 1-propanol

1.00 2.18 2.78 2.97 6.95 14.2 17.2

Capacity factor on LiChrosorb

Solute

0.521

0.473

86

0.92 1.95 2.53 2.70 6.40 12.5 15.1

0.641

0.776

141

0.86 1.81 2.41 2.56 5.96 11.3 13.7

0.856

2.58

470

0.64 1.31 1.92 1.97 4.22 7.37 8.94

0.301

0.187

34

0.635 1.26 1.87 1.94 4.12 8.43 10.22

0.410

0.302

55

0.595 1.17 1.71 1.80 3.91 7.81 9.45

0.487

0.413

75

0.58 1.13 1.60 1.68 3.78 7.43 9.00

Capacity factor on LiChrosorb

0.647

0.797

145

0.51 1.01 1.42 1.50 3.42 6.45 7.83

Si 100

0.716

1.10

199

0.48 0.935 1.33 1.40 3.21 5.90 7.16

VARIATIONS OF CAPACITY FACTOR VALUES OF TEST SOLUTES VERSUS WATER CONTENT IN 1,2-DICHLOROETHANE

TABLE III

0.947

2.41

438

0.40 0.76 1.15 1.20 2.58 4.53 5.51

0.220

Water mole fraction on free silanol groups, IXH201.

0.330

0.413

0.506

0.341

72.5

1.60 3.62 4.47 4.19 9.24 16.0 20.1

Si 60

0.628

0.198

Water mole fraction on free silanol iTOUPS~IXH*Ol.

0.312

0.482

0.664

0.564

0.266

0.129

0.0703

iXHIOlrn. lo3

85

40

19.5

10.6

Water content in mobile phase (ppm)

Si 60

14.8 11.3 4.09 11.6 12.15 11.3

on LiChrosorb

18.8 14.4 5.03 13.55 14.55 12.9

25.65 19.5 6.69 16.95 18.55 15.9

Aniline pNitroaniline o-Nitroaniline Phenol pCreso1 b-Napbthol

23.0 17.6 6.03 15.6 16.95 14.7

Capacity factor

Solute

0.733

0.783

118

12.8 9.86 3.62 10.6 11.0 10.25

0.357

0.438

0.474

0.300

63.5

55 0.260

0.99 1.97 2.77 2.58 5.54 9.62 12.0

1.02 2.01 2.86 2.66 5.63 9.88 12.2

Si 100

0.852

1.64

0.904

2.69

405

7.46 7.22 7.30

8.67 8.71 8.48 247

5.77 5.36 2.05

0.295

0.120

18

7.80 8.26 7.34

9.60 6.84 2.80

0.383

0.177

26.5

7.32 7.90 6.87

8.88 6.27 2.52

0.495

0.280

42

6.70 7.17 6.28

7.97 5.63 2.30

Capacity factor on LiChrosorb

WATER CONTENT IN CHLOROFORM

0.209

0.185

39

18.5 0.088

1.09 2.15 3.00 2.80 5.97 10.7 13.2

1.21 2.41 3.32 3.13 6.53 12.3 15.0

Capacity factor on LiChrosorb

8.73 7.11 2.77

0.793

1.28

270

119 0.563

1.16 2.47 3.43 3.30 6.63 10.5 13.4

1.46 3.30 4.11 3.88 8.41 14.1 17.8

VARIATIONS OF CAPACITY FACTOR VALUES OF TEST SOLUTES VERSUS

TABLE V

0.094

IXH*Olrn . lo3

0.235

49.5

35

20

Water content in mobile phase (ppm)

0.164

1.69 3.89 4.63 4.31 9.77 17.4 21.8

1.79 4.12 4.91 4.55 10.3 18.7 23.3

on LiChrosorb

1.91 4.48 5.20 4.81 11.0 20.4 25.4

Capacity j&for

o-Nitroaniline p-Nitroaniline p-Cresol Phenol 1-Phenyl- l-propanol I-Phenyl-2-propanol 3-Phenyl- 1-propanol

Sotute

0.561

0.690

0.636

0.743

0.826

124

1.65 5.30 5.52 5.09 1.80 5.71 5.97 5.45 96

5.80 3.96 6.23 4.37

0.776

1.15

245 90 0.426

0.72 1.41 2.15 2.05 4.06 6.39 8.12 0.92 1.82 2.61 2.44 5.21 8.86 11.0

Si 100

VARIATIONS OF CAPACITY FACTOR VALUES OF TEST SOLUTES VERSUS WATER CONTENT IN DICHLOROMETHANE

TABLE IV

INFLUENCE

OF WATER IN LSC SYSTEMS. II.

383

TABLE VI VARIATIONS OF CAPACITY FACTOR VALUES OF TEST SOLUTES VERSUS WATER CONTENT IN CARBON TETRACHLORIDE Solute

Capacity factor

on LiChrosorb

Si 60

Naphthalene o-Methylmethoxybenzene p-Methylmethoxybenzene Methoxybenzene o-Ethylnitrobenzene o-Methylnitrobenzene p-Methylnitrobenzene Nitrobenzene

1.20 6.68 14.7 11.05 6.18 8.56 12.2 9.99

0.98 5.36 11.55 8.73 5.29 7.20 10.2 8.28

Water content in mobile phase (ppm)

2.4

4.7

7.5

IXH201mIO4

0.205

0.402

0.642

0.856

2.35

Water mole fraction on free silanol groups,

0.321

0.48 1

0.596

0.663

0.844

0.825 4.70 9.85 7.32 4.74 6.38 9.04 7.12

0.715 3.85 7.98 6.07 4.26 5.63 7.89 6.29 10.0

0.43 1.76 3.72 2.97 2.73 3.49 4.99 4.07 27.5

bH20i.

We can also remark that selectivity is affected by the water content of the solvent. It is seen that, in a general way, the higher the water mole fraction on the free silanol groups, the smaller is the selectivity. Once again, from the foregoing, in LSC the water must be considered as a polar solvent and its content must be carefully controlled. As a rigorously anhydrous solvent does not exist, we always have a binary mixture: solvent plus water. Solute-solvent equilibrium constant values From eqns. 17 and 18 the equilibrium constants K1, K2 can be calculated. The TABLE VII VARIATIONS OF CAPACITY FACTOR VALUES OF TEST SOLUTES VERSUS TENT IN CYCLOHEXANE

WATER CON-

Solute

Capacity factor

Benzene Fluorobenzene Chlorobenzene Nonadecylbenzene Naphthalene

2.66 1.70 1.36 1.48 4.86

2.33 1.53 1.21 1.28 4.27

2.18 1.45 1.18 1.21 3.99

1.66 1.17 0.945 0.87 2.94

1.35 1.01 0.835 0.755 2.53

1.30 0.995 0.82 0.72 2.38

1.25 0.965 0.795 0.685 2.28

Water content in mobile phase (ppm)

1.8

2.8

3.3

5.9

8.1

8.9

9.4

IXH201rn. lo4

0.084

0.131

0.154

0.275

0.378

0.415

0.439

0.957

0.296

0.395

0.435

0.579

0.654

0.675

0.689

0.827

Water mole fraction on free silanol groups, IXH201.

on LiChrosorb

Si 60

0.74 0.67 0.57 0.305 1.19 20.5

C. SOUTEYRAND

384

0.0

I

0.30

0.40

I

I

0.50

0.60

1

0.70

8

0.30

et nl.

I

0.90

lXWh

Fig. 2. Variations of monolocalized solute capacity factors, k’, on LiChrosorb Si 60 versus water mole fraction, lxnZO18,on free silanol groups. Mobile phase: diisopropyl ether. Temperature: 25°C. Solutes: 0 = I-phenyl-1-propanol; 0 = I-phenyl-Zpropanol; n = 3-phenyl-1-propanol.

numbers of free (N) and bonded (IV’) silanol groups are knownl*; N,,, (number of solvent molecules contained in the dead volume of the column per gram of silica) is easily calculated from VO(dead volume of the column) and m (mass of silica in the chromatographic column) according to

where d, A4 and .N are the density, molecular mass of solvent and Avogadro’s number respectively. Thus, for each solute a and b in eqn. 18 are calculated by linear TABLE VIII WATER-CHLOROFORM EQUILIBRIUM CONSTANT VALUES CALCULATED FROM VARIATIONS OF SOLUTE CAPACITY FACTOR VALUES VERSUS WATER CONTENT IN THE CHLOROFORM Solute

K

Aniline p-Nitroaniline o-Nitroaniline Phenol p-Cresol j-Naphthol

3400 3300 3600 3400 3300 3500

Average value

3400

INFLUENCE

OF WATER IN LSC SYSTEMS. II.

385

k’

24

0.10

I

0.30

0.50

w

I

0.70

0.00

lXHzOla

Fig. 3. Variations of monolocalized solute capacity factors, k’, on LiChrosorb Si 60 versus water mole fraction, Ixn201.. on free silanol groups. Mobile phase: chloroform. Temperature: 25°C. Solutes: 0 = aniline; n = p-nitroanaline; q = o-nitroaniline.

4

I

0.10

0.30

I

0.50

I

0.70

I

0.00

Fig. 4. As Fig. 3 except test solutes: 0 = phenol; m = p-cresol; !J = p-naphthol.

*

I”n*ol.

C. SOUTEYRAND

386

ef al.

TABLE IX SOLUTECLASS N SOLVENT EQUILIBRIUM CONSTANT VALUES ON FREE SILANOL GROUPS (K,) OR ON BONDED AND FREE SILANOL GROUPS DEACTIVATED BY A MONOLAYER OF WATER (K2) FOR LICHROSORB Si 60 and Si 100 Solute

Solvent

Si 100

Si 60 4

Kl

Kl

Kl

Phenol

Diisopropyl ether 1,ZDichloroethane Dichloromethane Chloroform

2.7 26 55 160

1 11 22 36

3.1 34 68 155

1.2 10 21 44

Aniline

Diisopropyl ether Chloroform

36 280

3.2 32

33 210

3.9 40

o-Nitroaniline

Diisopropyl ether 1,2-Dichloroethane Dichloromethane Chloroform

p-Nitroaniline

Diisopropyl ether 1,2-Dichloroethane Dichloromethane Chloroform

38 24 60 215

6.8 7.7 16 25

37 22 55 160

7.7 7.9 14 26

I-Phenyl- 1-propanol

Diisopropyl ether 1,2-Dichloroethane Dichloromethane

6.6 68 140

1.8 28 43

6.4 70 140

2.1 28 47

I-Phenyl-2-propanol

Diisopropyl ether 1,ZDichloroethane Dichloromethane

26 165 300

5 45 77

25 160 355

5.5 46 80

3-Phenyl-1-propanol

Diisopropyl ether 1,2-Dichloroethane Dichloromethane

37 200 365

6.8 54 77

33 200 355

6.6 56 80

p-Cresol

1,2-Dichloromethane Dichloromethane Chloroform

29 63 170

12 22 35

35 71 165

11 23 45

/&Naphthol

Chloroform

145

37

140

43

Naphthalene

Carbon tetrachloride Cyclohexane

16 58

1.2 0.8

o-Methylmethoxybenzene p-Methylmethoxybenzene Methoxybenzene o-Ethylnitrobenzene o-Methylnitrobenzene p-Methylnitrobenzene Nitrobenzene

Carbon tetrachloride

89 205 155 67 100 145 125

6.2 8.7 6.6 12 14 19 13

Benzene Fluorobenzene Chlorobenzene Nonadecylbenzene

Cyclohexane

32 18 14 18

0.5 1.5 1.3 0.2

5.2 9.9 24 70

2.5 4.0 7.4 10

5.4 11 27 61

2.8 4.1 7.4 11

INFLUENCE

387

OF WATER IN LSC SYSTEMS. II.

regression and identified with their expression given by eqn. 17. The results are collected in Table IX. It must be pointed out, first that, within the experimental errors, the values of K1 and of Kz are identical on both silicas (Si 60 and Si 100). Secondly, K2 is always smaller than K1. This observation, again, demonstrates that the adsorption energy of the strong groups is higher than the adsorption energy on bonded groups or strong groups deactivated by a monolayer of water. Thirdly, for a given functional group, alcohol for instance, its equilibrium constant value and, consequently, its adsorption energy depends on the proximity and spatial volume of other groups. The influence of steric hindrance, due to a phenyl group, on the K1, Kz values of phenylpropanol position isomers is evident from Table IX: the closer the phenyl group is to the alcohol group, the smaller are the equilibrium constants K1 and K2. Relative

adsorption

energy of class N solvents

Considering the solute-solvent

equilibrium on free groups of silica, we can

write AC

KI tll-

-

e

--

RT

=

eAE

where AG is the net free energy and AE the net dimensionless free energy corresponding to eqn. 1 with:

AE = &,a + Em - Em - Es,, For class N solvents, the interaction energies for s and S with the mobile phase are cancelled by corresponding stationary phase interactions. The dimensionless free energy of adsorption is then:

AE = &,A - &,a Considering the system as ideal, we can write:

(22) In a similar manner: In K2 = ES,,, - Es+<

In K2 = Es., - ES,, From the K1 experimental values in Table IX, the reduced adsorption energy of class N solvents falls in the order: ~C2H4c12 a

>

,yCH+z a

>

E:‘-‘C’a

>

,57zc’4

>

E35H12

This classification is in good agreement with the solvent strength parameter, so, defined by Snyder and Glajch’ *. Conversely, the reduced adsorption energy of diisopropyl ether depends on the

C. SOUTEYRAND

388

et al.

solute nature. For example, E$SH14’ > Ec2H4C12for phenol and phenylpropanols as solutes and E3jH140 < E:2H4C12for p-nitroaniline as solute. This indicates that eqn. 22 is not valid with diisopropyl ether. This system must be considered as non-ideal. In that case, the equilibrium constants KI and K2 can be correlated with the thermodynamic equilibrium constant according to eqns. 10 and 11. Eqn. 22 can then be transformed into: In KI = ln ys.,,,+ Es,, - ES.. Consequently, with diisopropyl ether, the calculation of the solute adsorption energy involves the knowledge of the solute activity coefficient in this solvent. CONCLUSIONS

A simple retention model for monolocalized solutes and for class N solvents is proposed. This model gives a good agreement with experimental results for solvents having a reduced water content lower than 0.13, corresponding water mole fraction on free silanol groups of 0.75. In this range, water adsorption takes place only at free silanol groups. The solute molecule interaction with free, bonded and free silanol groups deactivated by a monolayer of water has been demonstrated. Water is considered as a polar solvent, and selectivity variations versus the water content of class N solvents are measured. It is seen that, in a general way, the higher the water mole fraction on free silanol groups, the smaller is the selectivity. From solute-solvent equilibrium constants, we have shown that the adsorption energy of a functional group depends on the proximity and spatial volume of other groups. REFERENCES 1 2 3 4 5 6 7 8

9 10 11 12 13

14 15 16

17 18

L. R. Snyder, Principles of Adsorption Chromatography, Marcel Dekker, New York, 1968. E. Soczewinski, Anal. Chem., 41 (1969) 179. L. R. Snyder, Anal. Chem., 46 (1974) 1384. L. R. Snyder and H. Poppe, J. Chromatogr., 184 (1980) 363. R. P. W. Scott and P. Kucera, Anal. Chem., 45 (1973) 749. R. P. W. Scott, J. Chromatogr. Sci., 18 (1980) 297. R. P. W. Scott and J. Traiman, J. Chromatogr., 196 (1980) 193. L. R. Snyder, in Cs. Horvath (Editor), High-Performance Liquid Chromatography, Vol. 3, Academic Press, New York, 1983, pp. 157-223. L. R. Snyder and T. C. Schunk, Anal. Chem., 54 (1982) 1764. L. R. Snyder, J. L. Glajch and J. J. Kirkland, J. Chromatogr., 218 (1981) 299. C. Souteyrand, M. Thibert, M. Caude and R. Rosset, J. Chromatogr., 262 (1983) 1. J. J. Kipling and E. H. M. Wright, J. Chem. Sot., London, 1 (1962) 855. G. Schay and L. G. Nagy, Acta Chim. (Budapest), 50 (1966) 207. S. K. Suri and V. Ramakrishna, Trans. Faraday Sot., 65 (1968) 1690. S. K. Suri and V. Ramakrishna, J. Phys. Chem., 72 (1968) 1555. S. Sircar and A. L. Myers, J. Phys. Chem., 74 (1970) 2828. J. Bizot, Bull. Sot. Chim. Fr., (1967) 151. L. R. Snyder and J.‘L. Glajch, J. Chromatogr., 214 (1981) 1.