PRACTICAL GUIDE

HPLC
A Practical Guide

JJJJJIPIPIP[

Contents
Chapter 1 Basic Concepts of High-performance Liquid Chromatography
1 2 3 4 5
Physical Parameters for High-speed Separations Physical Considerations Chemical Influences on the Separation Factor Basic Considerations of Liquid Chromatography Bibliography

1
1 4 5 6 9

Chapter 2 Instrumentation
Chromatographic Systems Injectors Eluent Delivery Systems Composition Gradient Delivery Systems Pressure Drop Changes due to Changes in the Eluent Composition Flow Rate Gradients Step-wise Elution Recycle Elution Column Switching Separation 4 Degassing Methods 5 Column Temperature Control 6 Detectors Spectroscopic Detectors Electrochemical Detectors Mass Spectrometer Other Detectors Measurement of Detector Sensitivity 7 Recording Systems Selection of Detector Time Constant 8 Columns and Connectors 9 Flow Cell Volume and Connecting Tube Dimensions for High Efficiency Operation 10 Other Components in a Liquid Chromatograph 11 Trou ble-shooting
1 2 3
vii

11
11 11 12 13

13 15 15 15 17 17 18 18 18 22 22 22 23 23 23 25 25 26 26

Vlll

...

Contents

12

References Reference Books

29 30

Chapter 3 Preparation, Testing, and Selectivity of Stationary Phase Materials

1

31
32 32 35 35 35 36 36 36 37 37 37 39 39 40 40 41 41 43 43 44 46 46 49 54 56

2 3 4

5
6

Synthesis of Stationary Phase Materials Silica Gel Surface Modification of Silica Gel Modification Using a Halogenated Silica Surface Modification Using a Chlorosilane Synthesis of Ion-exchanger Stationary Phases Cation-exchanger Anion-exchanger Organic Stationary Phase Materials Sieving of Stationary Phase Materials Column Packing Methods Column Evaluation Column Efficiency and Asymmetry Column Test for Reversed-phase Liquid Chromatography Inertness Test for Basic Compounds Inertness Test for Acidic Compounds Inertness Test for Heavy Metals Physicochemical Tests of the Stationary Phase Measurement of Void Volume Elution Volume of Proposed Void Volume Markers Selectivity of Stationary Phase Materials Silica Gel-based Stationary Phase Materials Organic Polymer-based Stationary Phase Materials Other Stationary Phase Materials References

Chapter 4 Selection of the Eluent

1

57
57 57
60
60

Reversed-phase Liquid Chromatography Selectivity of Organic Modifiers in the Eluent Comparison of the Effect of Organic Modifiers on Analytes with logP = 3 Chromatographic Behaviour of log P = 3 Compounds in Aqueous Tetrahydrofuran Chromatographic Behaviour of log P = 3 Compounds in Aqueous Methanol Chromatographic Behaviour of log P = 3 Compounds in Aqueous Acetone Chromatographic Behaviour of log P = 3 Compounds in Aqueous Dimethylformamide (DMF)

62 62

62

Contents

ix

Comparison of Effects of Modifiers on a Mixture of Analytes Chromatography of Ionic Compounds pH Effects Ion-pair Separations Theory of Ion-pair Liquid Chromatography Normal Phase Ion-pair Partition Liquid Chromatography Reversed-phase Ion-pair Liquid Chromatography pH Effects Effect of the Ionic Strength of the Buffer Solution Selectivity of Organic Modifier and the Concentration Effect Column Temperature Effects Detection Trouble-shooting in Ion-pair Liquid Chromatography Normal-phase Liquid Chromatography Classification of Solvents Preparation of the Eluent Size-exclusion Liquid Chromatography (SEC) Aqueous Phase Size-exclusion Organic Phase Size-exclusion References Bibliography of additional references on theoretical approaches in ion-pair liquid chromatography

62

65 65
70 70 71 72 75 80

80 80 80 81 81 85 89 90 91 92 93 95

Chapter 5 Separation Based on an Improved Column Efficiency

1 2

96
99 100 102 103 103 108

Improving Separation by Changing the Selectivity Improving Separation by Increasing the Column Efficiency Reducing the Eddy Diffusion Term Reducing the Effect of Longitudinal Diffusion Reducing the Effect of Resistance to Mass Transport in the Stationary and Mobile Phases Bibliography

Chapter 6 Influence of Physical Chemistry on Separations in Liquid Chromatography

1 Quantitative Structure-Retention Relationships in Reversed-phase Liquid Chromatography Prediction of Retention Times from log P in Reversed-phase Liquid Chroma tograp hy Prediction of Partition Coefficient (log P) Correlation between log P and log k Values Prediction of Retention Times from log P

109
109 110 110
110

111

Contents

4 5

Prediction of Retention Time of Ionic Compounds from log P and pK, Calculation of Dissociation Constant pK, from Hammets Equation Van der Waals Volume as the Basic Property Calculation of van der Waals Volume Prediction of Retention Times Based on van der Waals Volumes Enthalpy as a Thermodynamic Property in Retention Studies Measurement of Enthalpy by Liquid Chromatography Enthalpy Related to Retention Conclusion References

113 113 115 115 115 128 129 129 131 131

Subject Index

133

CHAPTER 1

Basic Concepts o Highf performance Liquid Chromatography

The two basic questions in high performance liquid chromatography focus on (a) how particular compounds can be separated, and (b) why particular compounds were separated by the liquid chromatographic method used. The answers can be obtained by the consideration of some simple representative chromatograms of the separation of well-known compounds. Such separations can be easily understood according to common principles of physics and chemistry . A separation is described by the following equation, which indicates the degree of resolution between two peaks in a chromatogram, R,. A complete separation requires R, > 1.2 units.
Rs

=&+

1 a-1

The resolution can be improved by increasing the column plate number, N , and/ or the separation factor, a (a = the ratio of the retention factors of the two compounds). N is the physical parameter and a is the chemical parameter for the separation. Higher N a n d a values give a better separation. The physical and chemical aspects of liquid chromatography, in addition to mechanical aspects, are briefly described in this chapter. Theoretical approaches are explained in detail in later chapters. The effect of stationary phase materials on the chemical selectivity is described in Chapter 3, and the influence of the eluent components is covered in Chapter 4. The plate number theory is discussed in Chapter 5. Quantitative optimization is explained in Chapter 6.

1 Physical Parameters for High-speed Separations

It was thought that high-speed separations would be achieved by the development of a physically stable pumping system and highly sensitive detectors;
1

Chapter 1

however, the main contribution to high-speed separation is made by small-size stationary phase materials. A shorter separation time with complete resolution cannot be achieved simply by increasing the flow rate or by using a small column. The theoretical plate number of a small column must be the same as that of a larger column to obtain the same separation. For example, the separation of a mixture of benzene, acetophenone, toluene, and naphthalene has been completed within 5.5 min using a 15 cm long, 4.6 mm i.d. column, packed with 10 pm porous octadecyl-bonded silica gel, whose theoretical plate number was 38 000 m- as shown in Figure 1.1A. Increasing the flow rate 4-fold reduced the separation time to 1.5 min, because this mixture was well separated (Figure 1.lB). The same mixture was separated within 4.5 min using a 10 cm long, 4.6 mm i.d. column packed with 3 pm octadecylbonded porous silica gel with a theoretical plate number of 117000 m- (Figure 1.1C). Doubling the flow rate resulted in completion of the separation within 2 min, as shown (Figure 1.1D). Comparison of these four chromatograms suggests that a fast separation can be performed either using a longer column with 10 pm stationary phase material with a high flow rate of the eluent, to give high resolution, or by a smaller

',

B
1 4

4 1 1

5 0 2 0 min

4 0 2

Figure 1.1 Eflect ofparticle size andflow rate on high-speedseparations. A and B column: 10 pm Cia-bonded silica gel, 15 cm x 4.6 mm i.d.; C and D column: 3 pm C18-bonded silica gel, 10 cm x 4.6 mm i.d.; eluent, 70% aqueous acetonitrile; flow rate A and C , 1, B, 4, and D, 2 ml min- I ; pressure drop A, 1.2; B, 5.9; C, 13.4; D, 27 MPa; detection: UV 254 nm. Peaks: 1, benzene; 2, acetophenone; 3, toluene, 4; naphthalene.

Basic Concepts of High-performance Liquid Chromatography

column packed with 3 pm stationary phase material. However, a high flow rate through the 3 pm stationary phase material is limited by a high column back pressure. The separation could also be completed within 1.2 min on the short column packed with 3 pm stationary phase material by using a stronger eluent, as shown in Figure 1.2. Furthermore, the sensitivity was also improved by using the smaller-size stationary phase material because the sample is less spread out in the eluent and is more concentrated when it reaches the detector. The actual peak height in Figure 1.1C is 1.6 times that in Figure 1.1A. A small column packed with small particle-size stationary phase materials promises high performance and a high-speed separation both in theory and in practice. The following equation describes the relationship of the column length (L) to the column efficiency: N = L/H. The high plate number N required for good separation is proportional to the longer column length L and small H value. The term H i s the height equivalent to a theoretical plate (HETP), which is the length of column needed to generate one theoretical plate. A good column has a high plate number for its length, and, thus, a good column has a low H value. The value of H can also be described by the following equation (which is described in detail in Chapter 5):

i 4

mln

Figure 1.2 Eflect on eluent component o f f o w rate. Column: 3 pm C18-bonded silica gel, 10 cm x 4.6 mm i.d.; eluent, 80% aqueous acetonitrile; f o w rate A, 1, B, 2 ml min-'; detection, UV 254 nm. Peaks: 1, benzene; 2, acetophenone; 3, toluene; 4,naphthalene. (Reproduced by permission from ref. 3)

Chapter I

This equation indicates that the particle size, dp,is the main contributor to the H value. The smaller the particles, the higher the theoretical plate number. The optimum condition is obtained by the relationship between the theoretical plate height and the flow velocity.

2 Physical Considerations
High-speed separations can be achieved with a short column packed with 3 pm stationary phase material, as shown in Figure 1.2. The sensitivity was also improved by the use of smaller-size stationary phase materials, due to less sample diffusion inside the column. The following conditions are required to obtain such a separation. a. small-diameter, spherical stationary phase materials that have high physical strength; b. a high pressure pump with controlled flow rate; c. a system that limits sample diffusion, by considering the column design, using small inner diameter connecting tubing, and a small volume detector flow cell; and d. a detector and recorder capable of a high-speed response. The theoretical plate number N of peak B can be calculated from the chromatogram given in Figure 1.3 by the following equation:
2

N = 16($)

where VR is the retention volume and w is the peak width at the base (measured in volume units). However, the retention volume includes the hold-up volume V , (also called dead volume). The hold-up volume is the sum of the void volume of the column ( V o = Y A ) , the volume of the injector ( O X ) and the volume of the detector and connecting tubing (XY) shown in Figure 1.3. The as actual separation efficiency is defined as the effective theoretical plate number N e ~which excludes the hold-up volume: ,
= 16(
VR -

Neff

VM

Commercial instruments have a reasonable balance between the recommended column size and the volume of the column and connecting tubing (XY). However, the theoretical plate number of a single column may give different values on different instruments, and even on replacement of the components and parts of a single instrument. Such discrepancies can be understood in terms of differences in the mechanics of the instruments and the design of their parts.

Basic Concepts o High-performance Liquid Chromatography f

Figure 1.3 A schematic chromatogram: VR, retention volume; V r , adjusted retention f volume; V, elution volume o peak; Vo, void volume; W ,peak width; V,, holdup volume; O X , volume of injector; X Y , volume of detector, including volume of tubing .

The normally acceptable extra-column dead volume (0Y in Figure 1.3) before there is a significant effect on the efficiency of a 15 cm long, 4.6 mm i.d. column should be less than 100 pl. This volume has to be reduced to less than 30 pl for a 5 cm long, 4.6 mm i.d. column. Replacement of the connecting tubing with shorter lengths of narrow-bore tubing and the selection of a smaller volume detector flow cell are necessary when using a shorter or narrower column. These changes together with a smaller column enable reductions in the volume of eluent required and in the separation time. This approach is economical and environmentally friendly. However, the reduced hold-up volume becomes technically very critical in the handling of smaller columns. Details of the basic mechanisms and the design of instrumentation are described in Chapter 2, which also covers the similarities and differences of various instruments.

3 Chemical Influences on the Separation Factor

One column can be used for different types of liquid chromatography by changing the eluent components. As an example, a column packed with octadecyl-bonded silica gel has been used for size-exclusion liquid chromatography with tetrahydrofuran (THF), normal-phase liquid chromatography with n-hexane, and reversed-phase liquid chromatography with aqueous acetonitrile. Examples of the chromatograms are shown in Figure 1.4. The elution volumes of polystyrene and benzene in the size-exclusion mode were 0.98 and 1.78 ml, respectively (Figure 1.4A). This means that separations by molecular size can be achieved between 0.98 and 1.78 ml in this system. In the normal phase mode the elution volumes of octylbenzene and benzene were 1.98 and 2.08 ml, respectively, in n-hexane solution (Figure 1.4B). This type of chromatography is called adsorption or non-aqueous reversed-phase liquid chromatography. These are adsorption liquid chromatography and nonaqueous reversed-phase liquid chromatography. The elution order of the alkylbenzenes in the reversed-phase mode using acetonitrile was reversed

Chapter I

A
E
5

B
4

C
1

C 0

>

c o (v

h1
I
0

12 1 3

I 5

I 0

I 5

I 0

I 5 min

Figure 1.4 D i e r e n t modes of chromatographs using the same column. Column, 5 pm C18-bonded silica gel, 15 cm x 4.6 mm i.d. Eluent A, tetrahydrofuran; B, nhexane; C, acetonitrile; flow rate, 0.5 ml min-' at ambient; detection, UV 260 nm. Peak 1, benzene; 2, ethylbenzene; 3 , butylbenzene; 4, octylbenzene; and 5 , polystyrene.

(Figure 1.4C). The elution volumes of benzene and butylbenzene were 2.01 and 2.52 ml, respectively. The elution volumes became larger with the addition of water to the acetonitrile eluent. In each case the elution orders are based on the solubility of the solutes (except in size-exclusion liquid chromatography). When separation cannot be achieved by improving the theoretical plate number of a column, it may be achieved by the selection of an appropriate stationary phase material and/or eluent. The degree of separation, the separation factor a, is the difference in retention volumes of analytes. The separation factor of two compounds is given by:

where VR2 and VRl are the retention volumes of peaks 2 and 1, respectively. These retention volumes depend on the properties of the solutes (analytes), stationary phase materials, and eluent components. A higher a value, i.e. an increase in the difference in retention volumes, can be achieved by using a different stationary phase material and/or eluent. Details of the selection of stationary phase materials and eluent are described in Chapters 3 and 4.

4 Basic Considerations of Liquid Chromatography

Identifying the most suitable separation conditions is the main objective of separation scientists. It is easier for skilled chromatographers, but this is a complicated subject for beginners. One approach is to find a chromatogram that exhibits the separation of a similar mixture. However, similar mixtures may have been separated under very different conditions; either the separation

Basic Concepts of High-performance Liquid Chromatography

columns and/or the components of the eluent may have been different. Furthermore, the elution order is sometimes reversed. When an appropriate chromatogram is found in the literature, the conditions may need to be modified to take into account the other compounds in the mixture, any necessity for sample pre-treatment and the purpose of the separation. The chromatographer should be familiar with the capabilities and requirements of the following methods. a. Pre-treatment of samples, stationary phase materials, and elution solvents; b. Separations based on molecular size; (i) aqueous size-exclusion liquid chromatography; (ii) non-aqueous size-exclusion liquid chromatography; c. Normal-phase liquid chromatography; d. Reversed-phase liquid chromatography; e. Ion-exchange liquid chromatography; f. Ion-pair liquid chromatography; g. Chiral separation and affinity liquid chromatography. Choosing the sample pre-treatment method is difficult. The most important consideration is the final condition of the target compounds. What kind of solution is obtained? The type of solvent and the concentration of the target compounds are very important in the selection of the separation conditions. The pre-treatment of stationary phase materials is also important for silica gel and ion-exchangers, even when a new column is being used. Pre-treatment of the elution solvent is also important. High-performance liquid chromatography grade solvents from different manufacturers contain different amounts of impurities. The purity of the water is especially critical. The specified solvent for a special preparation stage is often not compatible with the desired chromatography. A large amount of such a solvent should first be injected, followed by measurement of the background of the chromatogram. How are analytes retained on, or in, a stationary phase? This depends on the physicochemical interaction between the analytes and the stationary phase material. When a strong solvent, in which the solute readily dissolves, is used for elution the solute is eluted very quickly from the column. The forces holding an analyte on the stationary phase are similar to those responsible for dissolution in the solvent. Eight solubility properties are recognized: van der Waals (London dispersion) forces, dipole-dipole, ion-dipole, Coulombic and repulsion forces, charge-transfer complexation, and hydrogen-bonding and coordination bonds. The molecular interactions that are probably involved in retention in liquid chromatography can be explained by these interaction properties and are summarized in Table 1.1. The retention of a particular molecule is not due to only one property, but rather to a combination of several properties. The probable interactions can be estimated from the chemical structure of the analytes and stationary phase materials. The separation conditions employed for size-exclusion liquid chromatography are simple. A strong solvent for analytes and a suitable stationary phase

Chapter I

Table 1.1 CIassiJication in liquid chromatographic methods

Size-exclusion Reversed-phase Reversed-phaseion-pair Ion-pair partition Ion-exchange Normal-phase Charge-transfer Salting-out Ligand-exchange Chelation Affinity Chiral separation
phase material.

0 0 0 0 0 0

+ ooooo a + 000 + 0
000

000+n 0 0 0 + n

+oo

o +

00 0 0 0000 0 00 o + 00 o o + 0 0 0 0 0 0 0 0 0 000000 0

+
+

+ The most important mechanism, 0 important property, A interaction depending on stationary

material are necessary. If the impurities have high relative molecular masses ( M J , size-exclusion chromatography can be used effectively. Size-exclusion liquid chromatography in conjunction with a recycling system can also separate isomers; however, it is time-consuming and the columns are usually expensive. If a mixture of molecules with a M , of less than 2000 has to be separated and a recycling method seems to be insufficient for the separation, the following chromatographic technique can be carried out. If the sample concentration is large enough for chromatographic analysis, the eluted solution obtained by a size-exclusion chromatographic pre-treatment can be directly injected onto a liquid chromatograph using a syringe, after membrane filtration. If a good combination of stationary phase material and solvent cannot be found, then methods c-g in the above list are applied. In reversed-phase liquid chromatography, increasing the molecular size increases the hydrophobicity of solutes and results in a greater retention volume. This indicates that the van der Waals volume is an important property in optimization. Increasing the number of substituents with n-electrons and hydrogen bonding increases the solubility in water, that is they increase the polarity of the solutes. This indicates that dipole-dipole and hydrogen-bonding interactions contribute to hydrophobicity. Therefore, these properties are important in controlling the retention volume in reversed-phase liquid chromatography. However, the n-electrons of stationary phase materials such as

Basic Concepts of High-performance Liquid Chromatography

polystyrene gel and the hydrogen-bonding of non-endcapped bonded silica gels also contribute to the retention. Many compounds can be analysed by methods c or d, and sometimes both methods c and d are employed. For a preparative-scale separation, method c (normal-phase chromatography) is suitable due to the easy removal of solvent. Identifying the best separation conditions for ion-exchange liquid chromatography (method e) is a little more difficult. Therefore, if the compounds can be separated using methods c or d, these methods should be used. Even saccharides, organic acids, and nucleic acids are often separated by methods c and d. The separation speed in ion-exchange liquid chromatography is also slower than that in normal and reversed-phase liquid chromatography. Affinity liquid chromatography and chiral separations (enantiomer separations) require similar analyte properties. The solutes may have interactions through hydrogen-bonding, ligand formation, or Coulombic forces with the surface of stationary phase materials or the sites of additives; however, the selectivity is controlled by the steric effects of the structures of the analyte molecules and the recognition molecules (chiral selectors). The physical and chemical properties of stationary phase materials are described in Chapter 3 (including methods for their synthesis) to clarify the differences in similar stationary phase materials supplied from different manufacturers. A detailed selection guide to solvents is given in Chapter 4. The unlimited selection of eluent components and their concentrations is a powerful force in developing separations in liquid chromatography. Although this area seems rather complicated, it is easy to understand the selection of a suitable eluent when you first identify the molecular properties of the analytes and solvents.

5 Bibliography
J.C. Giddings, Dynamics of Chromatography 1. Principles and Theory, Marcel Dekker, New York, 1965. M.T. Gilbert, High-performance Liquid Chromatography, Wright, Bristol, 1987. S. Hara, S. Mori, and T. Hanai, Chromatography. The Separation System, Maruzen, Tokyo, 1981. H. Hatano and T. Hanai, New Experimental High Performance Liquid Chromatography, Kagaku Dojin, Kyoto, 1988. E. Heftmann, Chromatography, 5th edn, Journal of Chromatography Library, Vol. 5 1A, Part A Fundamentals and Techniques, Elsevier, Amsterdam, 1992. A.M. Krstulovic and P.R. Brown, Reversed-Phase High-Performance Liquid Chromatography, Wiley, New York, 1982. S. Lindsay and D. Kealey, High Performance Liquid Chromatography, Wiley, Chichester, 1987. V.R. Meyer, Practical High-Performance Liquid Chromatography, Wiley, Chichester, 1994. P.A. Sewell, B. Clarke, and D. Kealey, Chromatographic Separations, Wiley, Chichester, 1987. R.M. Smith, Retention and Selectivity in Liquid Chromatography, Journal of Chromatography Library, Elsevier, Amsterdam, 1995, Vol. 57.

10

Chapter I

L.R. Snyder and J.J. Kirkland, Introduction to Modern Liquid Chromatography, Wiley, New York, 1979. L.R. Snyder, J.J. Kirkland, and J.L. Glajch, Practical HPLC Method Development, 2nd edn, Wiley, New York, 1997. R.W. Yost, L.S. Ettre, and R.D. Conlon, Practical Liquid Chromatography, An Introduction, Perkin-Elmer, Norwalk, 1980.

CHAPTER 2

Instrumentation
The basic concepts of the instrumentation for liquid chromatography are described here, with the mechanism of their operation and their influence on the separation of analytes.

1 Chromatographic Systems
The components of all chromatographic systems are basically the same; however, the specifications and sizes differ between systems. A schematic diagram of a full-scale computerized system is shown in Figure 2.1. This type of system was originally proposed at the start of studies to design instrumentation for liquid chromatography, and the fully automated system is now commercially available. It is powerful, providing it is trouble-free. Its operation seems simple to the user, but trouble-shooting for this system is complicated. The basic concept of each of the components of the system is explained in the following sections.

2 Injectors
The maximum injection volume depends on the volume of the sample loop in the injection valve. The reproducibility of manual injection depends on the skill of the operator. The use of a small sample loop and an overflow injection of the sample solution so that the loop is fully flushed with sample are basic requirements for quantitative analysis. The highest injection reproducibility can be obtained by an auto-injector with a fixed sample loop. The smallest reasonable injection volume is 1 pl. A nl-scale injection valve can be constructed; however, the memory effect at the surface of contact parts affects quantitative analysis compared with the use of a pl-scale injection valve. For a semi-micro system, a low hold-up volume injection valve is desired. The minimum injection volume is 80 nl. For a preparative-scale injection, the sample loop can be easily replaced with a larger-volume loop, such as a 200 pl, instead of the standard 20 pl loop.
11

12

Chapter 2

.............................................

T2

Figure 2.1 Automated chromatograph. Components: solid line, liquid f l o w line; dotted line, communication line; S1, S2, and S3, solvent reservoirs; DG, degassing system; V1, V2, and V3, valves; Pu, pump: Pre, pressure sensor: Inj, injector; C , column; F,fraction collector; RC, reaction bath; RR, reaction reagent; D, detector; W, waste; T1 and T2, oven; COM, computer; Pro, solvent programmer; Pri, printer; INT, integrator or recorder.

3 Eluent Delivery Systems

The pump is the heart of HPLC and must satisfy a number of requirements. It must deliver a constant and pulse-free flow (stability: less than f0.5%, or usually less than & 1 pl min- I ) which must be reproducible under high pressures up to 50 MPa. The internal volume of the pump must be very small to enable rapid eluent changes and enable a precise gradient to be formed. The material of the flow lines must be compatible with all types of solvent, including organic solvents and strongly acidic and basic solutions. Four types of eluent delivery systems are popular. Widely used are the syringe pump and reciprocating piston pumps. A syringe pump is used for micro-scale liquid chromatography and

Instrumentation

13

capillary liquid/supercritical fluid chromatography. The most widely used is the double-piston reciprocating pump. This pump is suitable for a wide range of flow rates, from semi-micro liquid chromatography (LC) to preparative-scale LC under high pressure. The reciprocating piston pump design has been further modified in the masterlslave reciprocating pump; one piston works to pump the eluent and the second piston works as the flow controller. The degree of pulsation of the flow depends on the mechanical and electrical capability of the manufacturer. The latest models provide improved precision of the flow rate, and can handle semi-micro-scale liquid chromatography. A precise flow rate ( f 1 pl min- ') improves the reproducibility of gradient elution on standard size columns (6 mm > column i.d. > 3 mm). With the alternative diaphragm pump, the maximum pressure is less than that of the reciprocating pump, but the diaphragm resists a wider variety of solvents. A pneumatic amplifier pump is suitable for high flow rate operations such as preparative-scale liquid chromatography, but may not deliver a constant flow rate as it operates at constant pressure. This type of pump is also used for column packing. Several methods are applied to reduce the separation time. The best way is the selection of a suitable column and an eluent using isocratic elution. However, much skill and knowledge are required to make such a system. A flow rate gradient, step-wise elution, or eluent composition gradients are commonly applied to reduce separation times.

Composition Gradient Delivery Systems

High and low pressure gradient systems are available, which mix two eluents to give a defined increase in eluent strength through a separation run, effectively reducing the separation time. A low pressure gradient system requires the premixing of two or more solvents before pumping. This system is economical, but the internal system volume is higher than that of the high pressure gradient system. The precision of the high pressure gradient system, in which two pumps deliver the eluent components at changing rates to provide a total constant flow with a changing composition, is better but the cost is higher. The change and the reduction in elution time from an isocratic to a gradient elution are shown in Figures 2.2 and 2.3. The maintenance of a gradient elution system is more complicated than one using isocratic elution, but gradient elution is necessary for the separation of complex mixtures, such as amino acids, nucleic acids, and many biological samples.

Pressure Drop Changes due to Changes in the Eluent Composition

During a gradient elution the operational back-pressure changes due to changes in the viscosity as the eluent mixture changes' (Figure 2.4). When methanol and water are used as the components, the back pressure drop can become particularly high and sometimes stops the pump because it exceeds its maximum pressure limit. A high back pressure must be prevented, particularly when a longer column packed with small particle size is being used.

14
1

Chapter 2

I 0

I 10
min

I 20

30

Figure 2.2 Separation of aromatic compounds using isocratic elution. Conditions: column, 5 pm C18-bonded silica gel, 15 cm x 4.6 mm i.d.; eluent, 0.001 M phosphoric acid in 55% aqueous acetonitrile; flow rate, 1 ml min-'; temperature, ambient; detection, UV 254 nm. Peaks: 1, phenol; 2, 4-methylphenol; 3, 2,4dimethylphenol; 4, 2,3,5-trimethylphenol; 5 , benzene; 6, toluene; 7 , ethylbenzene; 8, propylbenzene; and 9, butylbenzene.

A 55 %

!
75 %
2

95 %

3
4 5 6 7 8 9

> 3

-.JL
I 0

I
10

min

Figure 2.3 Gradient elution of aromatic compounds. Chromatographic conditions as Figure 2.2 except f o r acetonitrile concentration.

Instrumentation

lm5
-

1.o

0.5

v
Methanohater Acetonitrilehater

15

-0 0 50 % 100 concentration of water in organic modifier

Figure 2.4

Relationship of viscosity to water and organic modifier mixture ratio

Flow Rate Gradients

Increasing the flow rate also reduces the separation time. However, this method is limited by the physical strength of the instrument and by the loss of detection sensitivity. The chromatograms of aromatic compounds separated isocratically and with a flow rate gradient are shown in Figures 2.2 and 2.5, respectively. Doubling the flow rate roughly halved the separation time; however, the peak areas became smaller. Flow programming elution is suitable for the separation of a mixture of homologous compounds.

Step-wise Elution
If the separation by isocratic elution is perfect, the eluent can be switched during the run to a stronger one to speed up late components (Figure 2.6). The peak height of longer-retained compounds increases and the retention time is shorter compared with the chromatogram of isocratic elution (Figure 2.2). This method is not suitable for quick repeat injections, due to the requirement for a significant column conditioning time after returning to the original conditions.

Recycle Elution
The recycle elution method can be applied to mixtures of very similar compounds that cannot be fully separated by a single pass through the column. This method makes more effective use of a column. The effluent from the column is repeatedly re-passed through the same column. The number of cycles multiplies the total theoretical plate number of the column if the system is

16
1

Chapter 2

2 mL

I
0

10 min

20

Figure 2.5 Flow rate programming elution of aromatic compounds. Experimental conditions as Figure 2.2 except the f l o w rate. Flow rate increased f r o m 1 to 2 ml min-'.
1

55r
75 %
I 0

I
10

min

Figure 2.6 Step-wise elution of aromatic compounds. Experimental conditions as in Figure 2.2 except the acetonitrile concentration was changed from 55 to 75% in water.

carefully assembled. This method is usually applied to larger columns and preparative-scale separations. Figure 2.7 shows the recycle chromatogram of the purification of a mixture of anthracycline antibiotics, in which a fraction (or heart-cut) of each separation is recycled through the column, becoming successively more resolved on each cycle.

Instrumentation

17

* the most active fraction

1st cycle

2nd cycle

3rd cycle

10 cycle

Figure 2 7 Purlfication of a mixture of anthracycline antibiotics using heart cut recycle . liquid chromatography to yield the most active fraction. Conditions: column, J A IGEL-3 10; eluent , chloroform-methanol-25 % N H 4 0H (200 :5 : ) ;Jlo w I rate, 4 ml min-'; detection: refractivity index. (Reproduced by permission from Japan Analytical Industry data)

Column Switching Separation

The selectivity of different stationary phase materials can be applied using columns in sequence to provide high-speed isocratic separations instead of gradient elution. An example for amino acids analysis is shown later in Figure 4.15, where the same eluent was used for all of the separations and the fraction containing the sample components of interest was switched from one column to another.

Degassing Methods

Degassing of the eluent is important for trouble-free operation and highly sensitive detectioq2 otherwise the eluent may become supersaturated with air that is released as bubbles in the pump check valves or the detector flow

18

Chapter 2

cell. Several methods have been proposed, and two on-line systems have been commercialized. One method is bubbling helium gas into the eluent. Another is an on-line vacuum. Both mechanisms are simple, but the latter system is popular due to its compact size, low hold-up volume, and because it does not require expensive helium gas. Even if the eluent is a premixed mixture of water and organic solvent, pre-degassing is still important. The eluent container should be degassed using a water pump vacuum. If the operation is performed in an ultrasonic bath, degassing is completed within 1 min. Longer degassing can cause loss of the more volatile eluent components.

5 Column Temperature Control

If good reproducibility of the retention times is required, the column temperature should be kept constant. A water bath or an air oven are common systems, and thermo-tape is also effective if the column is made of steel. When temperature control is extremely important, pre-heating of the eluent is necessary. If the whole apparatus is placed in a temperaturecontrolled box, the temperature is constant, but such a system is not usually required.

6 Detectors
One of the first steps in the development of modern liquid chromatography was the automation brought about by adding an on-line detector. Subsequently, the development of small particle size high-performance stationary phase materials provided more efficient separation systems, which then became known as high-pressure or high-performance liquid chromatography (HPLC). A variety of analytical instruments have been employed as detectors. Originally, standard instruments were modified to monitor effluents from a high-performance column. The highest possible sensitivity and selectivity of detectors are required in biomedical research, and led to the introduction of dedicated instruments with high qualitative performance. The basic mechanism is the same as that of the original instruments, but technical improvements have been made in the mechanics and electronics. The specifications of modern detectors for HPLC are summarized in Table 2.1.

Spectroscopic Detectors
Energy from a light beam is absorbed by molecules with a chromophore. An absorption spectrophotometer uses this mechanism, and the energy loss depends on the concentration and molecular absorption constant of analyte molecules and the wavelength of the light. The most popular detector, the

Instrumen tat ion

19

Table 2.1 Specijicity o detectors f

Detector Ultraviolet (visible) spectroscopy Sensitivity 1 ng ml-' 1 10-~ AUFS Cell volume Remarks Pl 4.5-20 Common detector; selection of wavelength improves the accuracy of quantitative analysis, wavelength scanning or photodiode array permits qualitative analysis. W-lamp or complex lamp is suitable for visible region detection Sensitive for fluorescent compounds All purpose; temperature control for high sensitive detection, usually used for preparative LC or size exclusion LC, eluent must be isocratic Specific for optical isomers Suitable for carbonyl compounds; NaC1, CaF2 and sapphire cells are available Detection limit 0.001 pmho is equivalent to 0.01 ppm NaCl, suitable for ion detection Suitable for ions, fatty acids, saccharides

Fluorescence spectroscopy Refractive Index

7-60 10 pg ml-' 1 x 10-4FS 5-14 10 ng ml-' 1.6 lo-' RIU

Polarimeter Infrared spectroscopy Conductivity

10 ng ml-

'

1.5

1 ng ml-I, 2 x 10-:pA 50 pg sCoulometry 5 x 10-8A Amperometry 50 ng ml-', 10 pA Pulsed 50 ng ml-', 10 pA amperometry Flame ionization I ngml-', 10 pA Electron capture 100 pg ml-'

Polarography

10

Suitable for catecholamines, phenols Suitable for saccharides All purpose; suitable for saccharides and non-volatile compounds Suitable for halogenated, phosphorus and oxygenated compounds All purpose; qualitative and quantitative detector Element selective Organometallic compounds
9 50-700

Mass spectrometer Inductive coupled plasma Atomic absorption Reaction heat Radioactivity

10 pg ml-' 20 ng mi-' 20 pg ml-' ng s - ' I4c2 n ~ i , : 3H :5nCi

All purpose Metabolite analysis

Refs: MS: Anal. Chem., 1981,53, 1603; ICP: J . Chromutogr. Sci., 1985, 23, 4;AAS, Analyst, 1981, 106, 921.

20

Chapter 2

ultraviolet absorption detector, uses a deuterium lamp as the light source. A visible region absorption detector requires a tungsten lamp. A combined lamp is useful for monitoring both in the ultraviolet and visible regions. Pre- or postcolumn derivatization of sample molecules to introduce a strong chromophore can be applied for highly sensitive and selective detection. Pre-column derivatization is also used to aid selective extraction from the effluent for collection purposes. Fluorescence detection is usually more sensitive than absorption detection, but the number of naturally fluorescent compounds is limited. Pre- or postcolumn derivatization can also be applied for this type of detection. Chemiluminescence detection is the most sensitive method for some fluorescent compounds. A light beam is refracted to different extents by different compounds. This mechanism is used for refractive index detection. This detector is not sensitive and the selectivity differences are negligible for homologous compounds, but any solvent with a different refractive index to the analyte can be used as the eluent. This detector is mainly applied to size-exclusion and preparative-scale liquid chromatography. Infrared spectroscopy is often used for qualitative analysis, and its powerful selectivity means that it can be used as a detector. However, the absorption of the eluent molecules, particularly in reversed-phase separations, often interferes with the detection of analytes. The infrared absorption detector therefore requires mechanical assistance to eliminate the solvent or needs powerful computer assistance to eliminate the background signal. Atomic absorption and inductively coupled plasma spectrometers are metalselective spectrometers used for organic metal analysis. The connection of these spectroscopic instruments to a liquid chromatograph is relatively simple. Chromatograms of alkylmercury3 and aminoplatinum analytes4 are shown in Figures 2.8 and 2.9, respectively. A polarimeter can be used for the identification of optically active comp o u n d ~as~shown in Figure 2.10. ,

I 0

I 10

min

I 20

Figure 2.8 Detection of alkylmercury compounds using flameless atomic absorption detector. Conditions: column, Corasil I , 50 cm x 2.1 m m i.d.; eluent, nhexane; flow rate, 0.5 ml min-'; detection, flameless atomic absorption spectrometer. Peaks: 1, benzene; 2, ethylmercury chloride; and 3, methylmercury chloride.

Instrumentation

21

cis

10 min

Figure 2.9 Inductively coupled plasma detection of diaminodichloroplatinum (11) . Conditions: column, Shodex O H , 25 cm x 4.1 mm i.d.; eluent, 0.01 M phosphoric acid; flow rate, 1 ml min-'; detector, emission at platinum line 265.9 nm. Samples: cis and trans diaminodichloroplatinums. The second peaks are considered to be their oligomers.

I 0

I 15 min

(-1

Figure 2.10 Chromatograms of ( f) -trans-stilbene oxide. Conditions: column, Chiralpak UP( +), 25 cm x 4.6 mm i.d.; eluent, methanol; column temperature, ambient;Jow rate, 1 ml min-'; detector A, polarimeter; B, refractive index.

22

Chapter 2

Electrochemical Detectors
Electrochemical detectors, which are based on the electrochemical oxidation or reduction of the analyte, can be applied to the analysis of selected compounds such as phenols. It is physically simple, but is very sensitive for catecholamines. However, the adsorption of reacted molecules on the surface of the electrodes can reduce the conductivity. To overcome this problem a pulsed voltage is applied, which cleans the electrode surface between measurements. This pulsed amperometric detection is also sensitive for carbohydrates. Conductivity detection, which involves measuring changes in the conductivity of an aqueous solution between two electrodes, is employed in ion chromatography for the detection of ionized analytes.

Mass Spectrometer
The mass spectrometer is a very sensitive and selective instrument. However, the introduction of the eluent into the vacuum chamber and the resulting significant pressure drop reduces the sensitivity. The gas exhaust power of a normal vacuum pump is some 10 ml min-' so high capacity or turbo vacuum pumps are usually needed. The gas-phase volume corresponding to 1 ml of liquid is 176 ml for n-hexane, 384 ml for ethanol, 429 ml for acetonitrile, 554 ml for methanol, and 1245 ml for water under standard conditions (O'C, 1 atmosphere). The elimination of the mobile phase solvent is therefore important, otherwise the expanding eluent will destroy the vacuum in the detector. Several methods to accomplish this have been developed. The commercialized interfaces are thermo-spray, moving-belt, electrospray ionization, ion-spray, and atmospheric pressure ionization. The influence of the eluent is very complex, and the modification of eluent components and the selection of an interface are therefore important. Micro-liquid chromatography is suitable for this detector, due to its very small flow rate (usually only 10 pl min-I).

Other Detectors
Flame ionization and electron capture detectors, which are used in gas chromatography, have been modified for liquid chromatography but have not been widely used. It is again necessary to eliminate the eluent to improve the sensitivity and to broaden their application. Moving-wire and moving-belt flame ionization detectors are useful for the detection of organic compounds, but a difference of greater than 80C between the boiling point of the analyte and the eluent is required to prevent loss of the analyte. The sensitivity of the electron capture detector is 0.12 ng ml- for aldrin; however, only volatile organic solvents can be used as the eluent components. A radioisotope detector has been used for radiolabelled metabolite analysis. Electron spin resonance (ESR) has been used for the analysis of radicals. Nuclear magnetic resonance spectroscopy (NMR) with a flow cell has also been used for qualitative analysis.

Instrumentation

23

The development of high-power NMR (800 MHz) spectrometers permits the simple operation of LC-NMR for metabolite analysis.

Measurement of Detector Sensitivity

For ultraviolet and visible spectroscopic detectors, a standard solution of a compound whose molar absorption constant is known must be prepared, and placed in the flow cell. The absorbance obtained is then compared with the value measured by a standard spectrophotometer. For a fluorescence detector, quinine sulfate is used as the standard compound. The flow cell is filled with a standard solution and the fluorescence intensity is measured. The value is compared with that measured by a fluorescence spectrophotometer. This standard solution is also used for fixing the wavelength and position of the flow cell. The Raman spectrum of water can also be used for this purpose. For refractive index detectors, 13.33 g of sucrose is dissolved in 100 ml of pure water and then diluted 200 times. The refractive index of the final solution RI, which is used to calibrate the instrument. is 1 x

7 Recording Systems
The visualization of the detector signal helps to clarify the separation conditions. Pen recorders were originally used, but the integrator is now popular due to the automatic reporting of both the retention time and peak area or height. The effective use of an integrator makes quantitative analysis easier. Computerbased integrators are powerful for the storage and further arrangement of data, and can also be used for column evaluation and as a system controller.

Selection of Detector Time Constant

The selection of the time constant of the detector and recording system is very important in high-speed separations. If the peak height monitored at a high flow rate is lower than that measured at the standard flow rate, what is the cause? Sometimes the peak height is lower than that expected theoretically. Carrying out a separation at a flow rate of 1 ml min-' and a recorder time constant of 1 s gave chromatogram A in Figure 2.1 1 , and at 4 ml min-' chromatogram B was obtained. The peak heights in chromatogram B are ca. 60% of those of A. The peak heights in chromatogram C, using a time constant of 2 s and flow rate of 4 ml min-', are ca. 50% of those of A. When the time constant was then reduced to 0.1 s, chromatogram D was obtained with peak heights ca. 80% of those of A. This experiment demonstrates the importance of time constant selection. The importance can be further observed by examining the peak height changes of the first peak in Figure 2.1 1. These chromatograms were monitored on a chart recorder. When the experiments were monitored by an integrator, the area of peak 1 in Figure 2.1 1 was constant in all four runs, but the peak heights

24

Chapter 2

B
4

2'

2 min

1 Figure 2.11 Eflect of time constant of detector. Conditions: column: 10 pm C 8 silica gel, 15 cm x 4.6 mm i.d.; eluent, 70% aqueous acetonitrile; flow rate for A, 1 ml min- ',for B-D, 4 ml min-I; time constant A and B, 1.O s, C , 2.0 s, D, 0.1 s. Peaks: 1, benzene; 2, acetophenone; 3, toluene; and 4,naphthalene.

! , 1
0 0

.2

+, I

1.0

sec

2.0

time constant

Figure 2.12 Ratio of peak area and height related to time constant of detector. Experimental conditions are the same as those in Figure 2.11. Peak height and area ratios are calculated from the data of peak nos. 1 and 4 as a % of the maximum value. Lines: A, peak height ratio of chromatograms measured at I .O ml min- pow rate; B, peak height ratio of chromatograms measured at 4.0 ml min-' flow rate; C, peak area ratio of chromatograms measured at 1.0 ml min-' flow rate; and D, peak area ratio of chromatograms measured at 4.0 ml min-'JEow rate.

'

Instrumentation

25

were affected by the time constant. The peak area ratio between peaks 1 and 4 was almost constant from 1 to 4 ml min-' (Figure 2.12, lines C and D), while the time constant was varied from 0.1 to 2 s. A rapid time constant is required when the flow rate exceeds 4 ml min-'. The peak height ratio of peaks 1 and 4 was affected by the time constant (Figure 2.12, lines A and B); therefore, the selection of the time constant for quantitative analysis is important. Furthermore, the optimum peak recognition parameters of an integrator should be determined for fast separations. When the time constant of the detector and the peak recognition parameter are too small, the noise level becomes high and quantitative analysis becomes difficult. Suitable conditions must be selected in each case for quantitative analysis.

8 Columns and Connectors

The best material for column tubing is glass, but careful handling is required due to its physical vulnerability. The most commonly used column material is stainless steel tubing, and the inner surface is polished like a mirror. These columns are physically stable and are compatible with a variety of compounds and eluent components. Some biological materials interact with metals; therefore, organic polymer and glass-lined stainless steel tubes are preferable for biological samples, and also for eluents with high ionic strengths which can attack the steel. The design of the inlet and outlet connections of columns depends on the physical strength of the materials. However, different connectors and fittings from different systems should not be mixed. Different connectors sometimes damage the efficiency of a whole system because they can create significant extra column dead volumes. The best approach is to use one connection system, even with instruments from different manufacturers. Replacement of the connection system is now simple, and does not damage individual components. Finger-tight polymer-type (PEEK) connectors are desirable if the instrument is operated at usual pressures (up to 20 MPa). Column systems are basically classified into two groups: conventional and cartridge columns. A conventional column usually demonstrates better column efficiency than does a cartridge column due to a higher column packing pressure, but its usefulness depends on the quality control system of the manufacturer and the skill of their technicians. A cartridge column is economical, but its theoretical performance and lifetime can be less than a conventional column.

9 Flow Cell Volume and Connecting Tube Dimensions for High Efficiency Operation
If the correctly sized flow cell and connecting tubing are not used, the high efficiency of a column or high theoretical plate number columns cannot be effectively used. The detector cell volume contributes hold-up volume. The larger is the cell volume, the greater the peak broadening. The cell volume

26

Chapter 2

should be less than 10% of the typical volume required to elute a peak. The peak volume is related to the retention factor k, and the longer the retention time, the larger is the peak volume. The cell volume is thus particularly critical for compounds with short retention times. A well-designed small volume flow cell is also required for smaller diameter or short columns. The relationship between the maximum desirable cell volume and column i.d. of a 10 cm long column is given by: Cell volume (pl) = 4.2 x 10-4x2 - 3.6 x 10-6x

+ 8.4 x

for a compound with k = 1, where x is the column i.d. The connecting tubing should be as short and narrow as possible. The volume of a 20 cm x 0.5 mm i.d. tube is 39.3 pl. That of a 20 cm x 0.25 mm i.d. tube is 9.8 pl, and that of a 20 cm x 0.125 mm i.d. tube is 2.5 pl. Some detectors are equipped with a heat exchanger that consists of a metal block containing a capillary tube. The volume of this tube also affects the theoretical plate number. If highly sensitive operation is not required, the heat exchanger can be removed or bypassed.

10 Other Components in a Liquid Chromatograph

A fraction collector and a post-column derivatization system were included (Figure 2.1) for a comprehensive and multi-purpose instrument. However, the fraction collector is needed only when collecting components from the effluent, and is generally not included in an analytical system. The post-column derivatization system is connected only when required for the selective and sensitive detection of specially targeted compounds. Usually, most compounds are directly detected by an on-line spectroscopic or other detector. A computer (Figure 2.1) can handle the complicated operations of controlling the auto-injector (injecting different volumes, adding an internal standard compound and selecting sample vials), the pumps (controlling the flow rate, selecting the different gradient methods, and changing the eluent composition), the detectors (selecting wavelength, sensitivity, time constant, and the optimization), the column temperature, the recording system, and other parts of the system. Additional software can improve the quality of reports, including the modification of chromatograms and the results of quantitative analysis. Furthermore, combination with a computational chemical analysis system can enable the production of quantitative structure-retention relationships.

11 Trouble-shooting
General problems and the appropriate maintenance procedures are summarized in Table 2.2. The details of individual instruments are well described in the manufacturers' manuals. The common problems are due to poor maintenance of the instruments and poor understanding of the specificity of stationary phase

Instrumentation

27

Table 2.2 Trouble shooting and maintenance

Trouble: Diagnosis 1. No response after switching on A. Fuse B. Electric cable Maintenance Replace fuse Check electric cable

2. No eluent flow from pump A. No movement of pump a. Nocurrent b. Lack of grease oil c. Safety pressure regulator on d. Gears jammed
e. Recrystallized salt B. Pump piston moves a. Leak in line b. Air bubble in pump c. Dirty check valve

See 1A and 1B Oil Fix or replace pressure gauge Change flow rate; if this does not work, then call the manufacturer Wash with water; call manufacturer Fix or replace appropriate part Degas the eluent, disconnect the pump and remove air by using a high flow rate Disconnect pump, wash with a high flow rate or remove check valve and wash in an ultrasonic bath with a strong solvent; finally, overhaul the check valve Fit new seal. Call manufacturer if using a syringe or diaphragm type pump Clean inlet side of column or replace filter

d. Aged piston seal C. Back pressure too high a. Plug in column, precolumn, guard column, or on-line filter due to dust from injector rotor seal or sample b. Expansion of stationary phase c. Solvent filter blocked

Wash column using intermediate solvent at low flow rate; change eluent; change chromatography system Replace with a suitable one

3. Eluent flow present A. Low flow rate a. 2Ba-d B. No response from detector a. No power to detector b. Aged lamp or photo cell c. Others C. No recorder movement a. No current b. Broken wire c. Malfunction of resistance wire d. Selection of polarity e. Flow cells unbalanced f. Others

lA, 1B Replace lamp or photo cell Call manufacturer lA, 1B Replace Clean; if this fails, go to f. Clean cell or replace solvent Use another recorder, and call manufacturer (continued)

28

Chapter 2

Table 2.2 Continued

Trouble: Diagnosis
4. Flow present and recorder responding A. Irregular noise on baseline a. Air bubble in flow cell b. Small bubble or immiscible solvent in flow cell c. Dirty flow cell window d. Insufficient light energy e. Malfunction of ground/earth B. Drifting baseline on output a. Unstable current or voltage b. Dust on light line c. Leak at flow cell d. Using refractive index detector e. Step-wise movement
Maintenance

Degas eluent; remove air bubble from flow cell Wash flow cell with ethanol, T H F or 6 M HN03 Apply b above; clean flow cell Replace lamp; if using an unsuitable eluent component of absorbance at detection wavelength, change eluent Fix or change earth connection Use stabilizer on power line Clean Fix it, or replace seal Flow rate too high; drafts; change prism p.osition Clean recorder slide wire; slightly increase gain of recorder Decrease recorder gain a little

f. Vibration of pen recorder C. Baseline drift without gradient elution Wash flow cell a. Dirty flow cell, expanding air bubble Control detector temperature b. Using refractive index detector Replace battery in recorder C. Aged reference battery of recorder Change chromatographic system d. Soluble stationary phase Ba e. Unstable current or voltage f. Expanding air bubble inside flow Remove air from line, especially from pump head; degas eluent cell D. Signal out of range of baseline 3B, C a. 3Bc 4A b,c b. 4A b,c Adjust flow cell position; remove flow cell c. Malfunction of light beam in and fix it detector Use HPLC grade solvent; use freshly d. Solvent prepared eluent Fill with the same eluent e. Reference flow cell Adjust zero balance f. Fresnel type refractive index detector 5. Bad peak reproducibility even when conditions 1-4 are satisfied A. Increased peak retention times a. 3A 3A Use fixed oven temperature b. Column temperature reduced If due to previous solvent, wash column c. Unstable column sufficiently, if due to stationary phase, wash column sufficiently; if stationary phase is soluble in the eluent, change chromatographic system d. Chart speed Fix recorder
~~

(continued)

Instrumentation

29

Table 2.2 Continued

~~~~

Trouble: Diagnosis

Maintenance

B. Decreased peak retention times a. Aged stationary phase b. Remains of previous eluent (especially gradient elution) c. Increasing column temperature C. Resolution becomes poor a. 5Ba-c b. Unbalanced density of column stationary phase c. Overloading of sample

Replace stationary phase or column; if stationary phase soluble, change chromatographic system Wash column sufficiently Fix oven temperature 5B a-c Repack column Resolution is poor even if repeated under initial conditions, see 5C, a,b Wash micro injection syringe; fix injector leakage 3B, b; 4A, c; 4C, b; 4D, f; if non-linear range of detector used, change concentration 3C, b,c; 4C, c; fix the gain of the recorder 5A Change chromatographic system

D. Poor peak area reproducibility

a. Poor injection b. Trouble with detector c. Trouble with recorder
6. Tailing peak A. Polluted column

7. Leading peak A. Unsuitable chromatographic system

materials and solvents. A high-performance liquid chromatograph may be used by many operators under different conditions. All operators should therefore study the maintenance manuals and keep a stock of frequently required replacement parts, such as rotor seals and pump piston seals. Many toxic organic solvents and eluent components are used in HPLC, and operators should therefore use safety glasses and work in areas that have adequate ventilation.

12 References
1. H. Colin, J. C. Diez-Mesa, G. Guiochon, T. Czajkowska, and I. Miedziak, J. Chromatogr., 1978, 167,41. 2. A. Nagai, in Advances in Liquid Chromatography, eds. T. Hanai and H. Hatano, World Science, Singapore, 1996, pp. 151-1 67. 3. W. Funasaka, T. Hanai, and K. Fujimura, J. Chromalogr. Sci., 1974, 12, 517. 4. T. Hanai, J. Hubert, G. Luissier, M. Bellavance, and M. Lefebvre, Proceedings of 4th International Symposium on Column Liquid Chromatography, Boston, May 1979. 5. T. Hanai, I. Suzuki, and K. Nakanishi, Anal. Sci., 1985, 1,483.

30

Chapter 2

Reference Books
J.W. Dolan and L.R. Snyder (eds.), Troubleshooting LC Systems, Humana Press, Clifton, 1989. E. Katz (ed.), Quantitative Analysis using Chromatographic Techniques, Wiley, Chichester, 1987. E.S. Yeung (ed.), Detectors for Liquid Chromatography, Wiley, New York, 1986. H. Parvez, M. Bastart-Malsot, S. Parvez, T. Nagatsu, and G. Carpentier (eds.), Electrochemical Detection in Medicine and Chemistry, VNU Science Press, Utrecht, 1987. R.W. Frei and K. Zech (eds.), Selective Sample Handling and Detection in HighPerformance Liquid Chromatography, Elsevier, Amsterdam, 1988. K. Blau and J.M. Halket (eds.), Handbook of Derivatives for Chromatography, 2nd edn, Wiley, Chichester, 1993. T. Hanai and H. Hatano (eds.), Advances in Liquid Chromatography, World Science, Singapore, 1996.

CHAPTER 3

Preparation, Testing, and Selectivity of Stationary Phase Materials

The selection of the stationary phase material is generally not difficult when the retention mechanism of the intended separation is understood. Stationary phase materials have different physical and chemical properties that can be understood from their chemical structure. They can be classified according to their retention mechanism and chemical structure (Table 3.1). The symbol indicates a strong interaction. All stationary phase materials can be classified in this manner, including chiral and affinity phases. For example, alkyl- or phenyl- bonded vinyl alcohol gels are used in reversed-phase liquid chromatography, and ion-exchange group-bonded vinyl alcohol gels can be used as cation or anion-exchangers. Non- bonded materials can be used for size-exclusion liquid chromatography. The fundamental behaviour of stationary phase materials is related to their solubility-interaction properties. A hydrophobic phase acts as a partner to a hydrophobic interaction. An ionic phase acts as a partner for ion-ion interactions, and surface metal ions as a partner for ligand complex formation. A chiral phase partners chiral recognition, and specific three-dimensional phases partner affinity interactions. Stationary phase materials are synthesized from different raw materials. Those stationary phase materials synthesized from inorganic materials, such as silica and alumina, are physically strong but chemically unstable. Conversely, stationary phase materials synthesized from organic materials, such as polystyrene or poly(viny1 alcohol), are chemically stable but physically weaker. Improvements in the chemical stability of inorganic stationary phase materials and in the physical strength of organic stationary phase materials are required; the marketed products do not have both and have to be used under restricted conditions in liquid chromatography.

31

32

Chapter 3

Table 3.1 Stationary phase material selection guide

Stationary phase material

Retention mode

Normal-phase LC Chemisorption, chelate Charge transfer Physical adsorption, hydrogen bonding Reversed-phase LC Ion-exchange LC Size-exclusion LC Affinity, optical isomer Hydrogen bonding acceptor Hydrogen bonding donor

a 0

++OOA
b + +

00

00000000000

++

vvv

4: strong interaction; 0: possible interaction, a: weak interaction; V: a wide variety of stationary phase materials available; a Ferrous ion in silica; especially COOH form of ion-exchanger.

1 Synthesis of Stationary Phase Materials

The selectivity of stationary phase materials can be understood if the method of their synthesis is understood. Differences in the same type of stationary phase material from different manufacturers or even from the same manufacturer depend on the synthetic methods and the quality control that has been employed. Details of the individual synthetic processes from different manufacturers have not been published, but are basically the same.

Silica Gel
Silica gel is a three-dimensional polymer of silicic acid, usually synthesized from tetrachlorosilane or sodium silicic acid. The reaction is as follows: either or SiC14 + 2H20
+

+ 2Ag20 -+4AgC1 + Si(OH)4

ion-exchange
+

NaSi04 + 4(H -form ion-exchanger) Si(OH)4 + 4(Na -form ion-exchanger)

Preparation, Testing, and Selectivity of Stationary Phase Materials

33

The removal of sodium and other metal ions is very important to obtain pure silica gels. The next step in both routes is: 2Si(OH)4 + (OH)3SiOSi(OH)3 + H 2 0 + silica gel Silica gel is also made from colloidal silica (silica sol): Si(OH)26- -+Si(OH)4 + silica gel The particle diameter and pore size of the silica gel materials depend on the concentration of silicic acid, the pH, the solvent and mixing conditions, and the reaction temperature. Technical skill is necessary for reproducible large-scale synthesis. The purity of silica gel made from sodium silicate depends on the ionexchange stage, and is usually about 99%. A purer silica gel can be produced from organosilica compounds. In this method tetraethoxysilane is polymerized to polyethoxysilane. The emulsion of polyethoxysilane in a water-alcohol mixture is further dehydroxylated and polymerized with a catalyst. The hydrolysed silica gel is sieved, dried, and sieved again before use as a stationary phase material. The pore size and particle size depend on the concentration of polyethoxysilane, the solvent, the temperature, the amount of catalyst, and the mixing conditions. Strong agitation leads to a smaller particle size, and a large amount of catalyst leads to a large pore size. The purity of this silica gel is more than 99.99%, depending on the purity of the water used for washing and sieving. The surface of the virgin silica gel is covered with water. Heating to 110 C forms a monolayer of silanol groups on the surface (8 pmol of silanol groups per m). Heating to over 600C produces a surface siloxane structure, and this is rehydrated to the silanol form under humid conditions, as shown in Figure 3.1. Electron microscopy photographs of classic irregular (A) and spherical (B) silica gels and high purity, spherical silica gel (C),whose purity is over 99.99%, are shown in Figure 3.2. The surface of pure silica gel is very smooth and physically very stable due to the homogeneous structure. Modification of the

-600C
H20

-.
\

Si

Si

) .
\
/O

Si

Figure 3.1

Modification o silica gel surface. f

34

Chapter 3

Preparation, Testing, and Selectivity of Stationary Phase Materials

35

surface is easier (see section below on surface modification) than for ordinary silica gels.

Surface Mod8cation of Silica Gel

Spherical porous silica gel is the easiest stationary phase material to handle; however, although it is physically strong it is chemically unstable. Surface modification can expand its capability for different modes of chromatography, such as normal-phase, reversed-phase, size-exclusion, and ion-exchange liquid chromatography. These stable modifications are performed by chemical derivatization of the surface silanol groups.
ModiJication Using a Halogenated Silica Surface. Silica gel is heated in 2 M HC1, filtered off, and washed with water; the dried silica gel is then reacted with pure thionyl chloride. After removal of the excess of thionyl chloride and byproducts (SO2 and HCl) under vacuum, chlorinated silica gel (FSiCl) is obtained. This is unstable in water and alcohol. The chlorinated silica gel can then be alkylated by one of a number of different reactions:

-Si-OH + (SOCl2, Tic&, or SiC14) + -Sic1 -Sic1 + (RMgC1) -+ r S i R + MgC12 (Grignard reaction) -Sic1 -+(RLi) -+ -SIR + LiCl -Sic1 + (RNH2) + r S i N H R + HC1 The chlorinated silica gel suspended in diethyl ether, dimethyl sulfoxide, or dioxaiie will also react with diamine or amino compounds. After being washed in an excess of amine and HC1, and then dried under vacuum, the final stationary phase materials are obtained: =Sic1 + H2N-(CH2),-X -+ -Si-NH-(CH2),-X X = CH3, NH2, C02H, S03H, CN, or NO2

+ HCl

Further modification of diamino bonded products with halogenated compounds can lead to a variety of stationary phase materials: -Si-NH-( CH2),-N H2 + Y-CH2-R-X R = CH2, C6H4
-+ =Si-N

H-( CH2),-NH-CH2-R-X

Modification Using a Chlorosilane. Silica gel is heated in 2 M HCl, filtered off, and washed with water. The dried silica gel is then boiled under reflux in toluene

Figure 3.2 Electron microscopic photos of silica gels: A, irregularly shaped; B, ordinary spherical; and C, high purity silica. (Contributed by Hideyuki Negishi, Kanagawa Dental College, Yokosuka, Japan)

36

Chapter 3

Figure 3.3 Molecular model of butyl-bonded silica gel.

containing the chlorosilyl reagent, using pyridine as the catalyst. Three types of chlorosilyl reagents are available: 3 ESi-OH 2 GSi-OH -Si-OH

+ RSiC13+ =SiO3SiR + 3HC1 + R2SiC12+ rSi02SiR2 + 2HCl + R3SiC1+ -SiOSiR3 + HC1

A depiction of a butyl-bonded phase is shown in Figure 3.3. Monochlorosilane reagents produce only monomeric phases; however, trichlorosilane reagents can produce both monomeric and polymeric phases depending upon the concentration of silyl reagent and the surface area. It is difficult to make a multilayer bonded phase, even when a large quantity of trichlorosilane is used for the reaction.

Synthesis of Ion-exchanger Stationary Phases

Cation-exchanger. Phenyl-bonded silica gel is suspended in chlorosulfonic acidacetone and boiled under reflux. After filtration and washing with acetone and dilute HC1, the sulfonated gel is then converted into a suitable salt form:

Anion-exchanger. Phenyl-bonded silica gel is suspended in chloromethyl methyl ether and boiled under reflux. Then, chloromethyl methyl ether containing anhydrous ZnC12 is added and the mixture boiled under reflux. After filtration,

Preparation, Testing, and Selectivity of Stationary Phase Materials

37

the gel is washed using 1,4-dioxane and diethyl ether. The washed silica gel is suspended in trimethylamine- 1,4-dioxane solution, and the mixture stored at 0 "C to give the quaternary ammonium ion exchanger: ESi-Ph
+ -Si-Ph-N+Me3

C1-

Organic Stationary Phase Materials

A variety of organic polymer gels are available for use as stationary phase materials. Their physical strengths and hydrophobicity depend on the properties of the monomers. Dextran gels are polar and soft. Vinyl alcohol copolymer gels are fairly polar. Polyacryl gels are physically hard and relatively non-polar. Polystyrene gels (a copolymer of styrene and divinylbenzene) are physically very strong and very hydrophobic. A higher divinylbenzene concentration makes a physically stronger polystyrene gel. The particle size, pore size, physical strength, and wettability in solvent depends on the reaction solvent, temperature, mixing conditions, and catalyst. Furthermore, ion-exchange groups can be introduced on the polystyrene surface to make ion-exchange resins.

2 Sieving of Stationary Phase Materials

The narrower the particle size distribution, the higher in theory is the potential theoretical plate number. A rough sieving is achieved by a water flow, air flow, or a vibration method. A common sieving method is Hamiltonian water flow (Figure 3.4). The particle distribution can be controlled within 1 pm by this method. A slurry of stationary phase material is allowed to float in the cylinder, and a solvent flows from the bottom to the top. The smaller and lighter particles float to the top of the cylinder and the larger and heavier particles sink to the bottom. The required particles are collected at the top of the cylinder. The selection of suspension solvent and control of the temperature are important.

3 Column Packing Methods

Dry packing is usually only used for large particle-size stationary phase materials, especially in preparative-scale columns. High-performance analytical-scale columns are packed by the wet slurry method. Balanced or viscous slurry methods are applied to reduce the sedimentation of the stationary phase materials during the packing process. Previously, balanced density slurries were made by mixing high-density halogenated solvents. However, halogenated solvents are toxic, so the viscous slurry method has been developed, which uses an ethylene glycol-alcohol suspension mixture. First, an empty analytical column, a pre-column, and a slurry reservoir are connected in series. The narrow-bore analytical column and pre-column are

38
suspension solution

Chapter 3

/
collection

Water flow

+
1

Figure 3.4 Hamiltonian particle sizing system.

filled with a packing solvent such as methanol. The reservoir is filled with the stationary phase slurry, and then the packing solvent is quickly pumped through the system under high pressure, depending on the physical strength of the stationary phase material. For example: for the preparation of a 15 cm long, 4.6 mm i.d. stainless tube column, 2.5 g of octadecyl-bonded silica gel was suspended in 25 ml of hexanolmethanol mixture, and kept in an ultrasonic bath for a few minutes to remove air. After the reservoir was filled with the slurry, methanol was pumped in at 10 ml min- under constant pressure, 45 MPa (450 bar). After the replacement of slurry solvent by methanol, the flow was stopped and the pressure allowed to drop. When OMPa was reached the reservoir was removed. Then, 20ml of water was added and methanol was again pumped in under the same conditions as before. Again, the flow was stopped and the pressure allowed to drop until it reached 0 MPa. The pre-column was removed and the analytical column closed. The maximum pressure that can be applied in the filling stage is based on the pore size, particle shape, and purity of the silica gel. This reproducible packing procedure is performed at constant temperature by using a water bath (6080 "C).

Preparation, Testing, and Selectivity of Stationary Phase Materials

39

4 Column Evaluation
The performance of columns is determined by using a simple procedure for both new and used columns. The test method and the results are usually printed on a sheet provided by the column manufacturer. Each manufacturer uses the most suitable method for their column to obtain the best theoretical plate number. The column evaluation provides important information for all users of liquid chromatography.

Column Efficiency and Asymmetry

Common standard compounds for reversed phase columns are toluene and naphthalene, which have retention factors, k , of about 3. The eluent modifier is methanol or acetonitrile at a concentration of 50-80%, depending on the hydrophobicity of the stationary phase material. For other stationary phase materials, corresponding analytes, with k = 3-5, can be used. The efficiency of a well packed column should be about 95 000 m-' 5% for a 5 pm octadecyl-bonded silica gel column. The equations for the calculation of the number of theoretical plates (N)and peak asymmetry ( A s ) are as follows. A model chromatogram is given in Figure 3.5. The value of N is given by:

where the retention time ( t R ) is 5.87 min in Figure 3.5. Assuming the peak is a Gaussian curve, the peak width at the base is considered to be 40 where 0 is the uncertainty in the retention times of individual analyte molecules. The peak width (in the figure) at half-peak-height (w+)is 0.17 min; and wb, the peak width at the base line, is 0.30 min. Therefore, N is 6605 from w;, and 6126 from wb. The difference is due to the uncertainties in the measurement of the peak width. For comparison between columns of different lengths the where L is the length of efficiency is often expressed as plates per m [ 1OO(N/L)], the column in cm. The value of the peak asymmetry As is calculated from: As
=

b/a

where a and b are, respectively, the widths of the front and back of a peak measured at 10% peak height from the baseline. The a and b values (Figure 3.5) are 0.12 and 0.16 min, respectively. Therefore, As is 1.33. The desired As value is 1.0-1.25 for good separations and a long column life. The lifetime of columns with As < 1.O is usually short, due to an inhomogeneously packed bed.

40

Chapter 3

V, = 1.59 O

J'
\ = 5.07 4
I
/ . I. \ .

I 6

min

Figure 3.5 Schematic of a typical chromatogram. V,, void volume; w+,width at halfpeak f height; wb, width at base; lOh, peak height; a, width o front at l h ; b, width of back at l h ; wi = 20.10h.

Column Test for Reversed-phase Liquid Chromatography Inertness Test for Basic Compounds
A bonded stationary phase whose silanol groups are completely modified by alkylation is, in theory, inert towards basic compounds. The lifetime of such columns is long, and can be maintained for more than 2000 h in 1% trifluoroacetic acid solution (pH 2) or 0.05 M disodium phosphate solution (pH 8.6). Such a column will withstand the passing of more than 10 litres of trifluoroacetic acid solution for continuous peptide separation. The peak asymmetry of basic compounds is excellent. The extent of unreacted silanol groups can be measured by the chromatographic behaviour of pyridine compared with The relationship of the different surface activity to the number of silanol groups is shown in Figure 3.6. If no active silanol groups are present, the

Preparation, Testing, and Selectivity of Stationary Phase Materials

41

A
1 !

B
2

C
2

I
0

I 10

I 0

I I 4 0

h
I 6

min

Figure 3 6 .

Test f o r active silanols on octadecyl-bonded silica gels. Column: 5 pm octadecyl-bonded silica gel, 15 cm x 4.6 mm i.d.; eluent, 60% aqueous acetonitrile; flow rate, 1 ml mintemperature, ambient; detection, UV 254 nm. Peak 1, pyridine; 2, phenol; and 3, toluene. A, Column with no active silanol groups; B, some active groups; and C , numerous active groups.

';

pyridine peak is sharp and its symmetry is like that in chromatogram A. Chromatogram B indicates the existence of some active silanol groups. The pyridine is retained longer than phenol and the peak shape is poor. In chromatogram C pyridine was completely adsorbed and could not be observed. The lifetimes of columns B and C were short in buffered solution. These columns are not suitable for the chromatography of nitrogen-containing basic compounds due to a long tailing peak or the disappearance of the peak by strong adsorption.

Inertness Test for Acidic Compounds

The retention and the peak asymmetry of benzoic acid also indicate the inertness of the bonded phase. If basic compounds remain on the surface or are used as reagents, the peak asymmetry of benzoic acid is poor. The peak height is lower than that of the same quantity of o-toluic acid.3i4 This phenomenon is observed if the basic catalyst that was used in the synthesis process has not been completely washed off the stationary phase or if active amino groups remain. This type of column is not suitable for the separation of acidic compounds.

Inertness Test f o r Heavy Metals

Active compounds often form a chelate with heavy metal ions, and elution from a column containing heavy metal ions is poor.3y4 existence of heavy metals The

42

Chapter 3

on the surface can be examined by using 8-hydroxyquinoline, which is used for heavy metal extraction from water. However, 8-hydroxyquinoline is regarded as a human carcinogen. Therefore it has to be used under controlled conditions. Other chelation reagents, such as pentan-2,4-dione and 2,3-dihydroxynaphthalene, can be used for the trace metal test, but they are not as sensitive as 8-hydroxyquinoline. The metal contents of different grades of silica gels are summarized in Table 3.2. Extra pure and pure silica gel contain only trace amounts of heavy metals but contain traces of alkaline metals from the distilled water used mainly during the sizing procedure. The effect of heavy metals in relation to the purity of silica gel is shown in Figure 3.7. The purer the silica gel, the sharper the peak shape, as in chromatogram A. From the peak tailing, Chromatogram B indicates the existence of heavy metals on the surface. The

Table 3.2 Metal content of silica gels of diferent purity

Silica gel Ordinary Purified Pure Extra pure
a Particle

PS*"
(pm)

p ~ * SA*' ~ (A) (m2g-') 122 120 100 100 333 350 350 450

Metallppm Na Mg 190 5 6 5

A1

Ca

Ti

Fe 22 50 13 1

Zn

5 5 10 5

250 3 nd 1

150 730 160 120 10 130 63 nd 70 5 9 n d

19 nd 6 1

size; pore size; surface area; nd, not detected by inductive coupled plasma spectroscopy (ICP), Cr, Mn, and Ni were not detected.

10

elution time

30 0 min

30

Figure 3.7

Test for trace metal ions on the surface of octadecyl-bonded silica gel. Column A, 5 pm octadecyl-bonded silica gel, 15 cm x 4.6 mm i d . ; By octadecylbonded silica gel, 15 c m x 4.6 mm i.d.; C, octadecyl-bonded 10 pm silica gel, 30 cm x 4.0 mm i.d.; eluent, 50% aqueous acetonitrile; flow rate, 1 ml m i n - ' ; temperature, ambient; detection, UV 220 nm (the scales of A, B, and C are not the same); sample, 8-hydroxyquinoline. Columns; A, metal ion f free; B, low proportion o metal ions; C , high level of metal ions.

Preparation, Testing, and Selectivity of Stationary Phase Materials

43

peak of 8-hydroxyquinoline could not be observed on chromatogram C. This also indicates that the surface treatment was poor. Pure silica gels (purity > 99.99%) are necessary for the separation of porphyrins. For the separation of biologically active compounds, bonded phases made from pure silica gels whose purity is known and guaranteed should be used.

Physicochemical Tests of the Stationary Phase

The carbon content of a stationary phase is measured by an elemental analyser, as a weight balance before and after heating at 800C. Particle size, pore size, and surface area are measured by specific instruments, such as a particle size analyser, nitrogen adsorption porosimeter, and mercury depression analyser, respectively. The precision of the measurement of carbon content is high; however, that of the other measurements is relatively poor. Therefore, it is difficult to relate the surface area of different silica gels to analyte retention factors.

5 Measurement of Void Volume

The retention factor, k, is the basic value in chromatography, and is related to the void volume (dead volume). The void volume is the space inside the column, where no retention of solutes has occurred and can be measured on a chromatogram, as shown in Figure 1.3. The void volume is about half the total volume of the column when it is packed with porous stationary phase materials. In practice, the effective void experienced by the analyte is smaller because the molecular mass of the analyte is usually much greater than that of the eluent molecule. In a model of porous stationary phase material, the pores can be represented as V-shape valleys (Figure 3.8), where region a is a support, such as

b
C

Figure 3.8 A, Schematic structure o porous stationary phase; a, support; b, bonded f phase; and c, adsorbed solvent layer. B, detailed view of eluent components in surface layer; , lower phase molecules; 0, upper phase molecules.

44

Chapter 3

silica gel or polystyrene gel, b is the bonded phase, such as octadecyl or ionexchange groups, and c is the solvent layer, which is the adsorbed components of the eluent. In Figure 3.8, if a sample molecule is small it can pass to the bottom of the valley. However, a larger sample molecule, such as a protein, is large, and jumps the valley if the pore size is small. The real void volume for a particular analyte thus depends upon its shape and mass. The pore size distribution of the stationary phase materials also affects the void volume of individual analytes. The exact value cannot be measured. A practical compromise method is examined below. In normal-phase liquid chromatography, the elution volume of the solvent is nearly constant in eluents having different mixtures of solvent components. The solvent peak can therefore be used to measure the void volume. In reversedphase liquid chromatography, the thickness of the organic solvent layer, which is c in Figure 3.8A, depends on the concentration of organic modifier and additives. In addition, an ion-exclusion effect is observed for ionized solutes, and the elution volume is often less than the exclusion limit of the column. However, measurement of the void volume is necessary to calculate k. Figure 3.8B indicates the microscopic interface of a partition. The symbols are lowerphase molecules and o are upper-phase molecules in a solvent partition. The distribution between the phases b and c is not clear-cut, as in a partition model. Phase b incorporates the bonded phase and the concentration of the organic component of the eluent is higher than in the bulk eluent; it gradually reduces towards the mobile phase in reversed-phase liquid chromatography. The effective void volume is therefore not constant.

Elution Volume of Proposed Void Volume Markers

Several compounds have been proposed for the measurement of the void volume, including sodium nitrate solution, water, deuterium oxide, fructose, acetonitrile, tetrahydrofuran (THF), meso-erythritol, gluconolactone, and 2,4dinitronaphthol. The elution volume of a number of these compounds has been measured in 10-90% aqueous acetonitrile and acidic-aqueous acetonitrile. The results are given in Figure 3.9 where the volumes in A and B were measured in aqueous acetonitrile and in C and D were measured in aqueous acetonitrile containing 50 mM phosphoric acid. Methanol (a) and deuterium oxide (g) showed two peaks when monitored by a refractive index detector (Figure 3.9C). Even the elution volume of acetonitrile was not constant in these eluents and varied from 1.25 to 1.95 ml. The volumes of ionized sodium nitrate (e) and 2,4dinitronaphthol (d) were smaller than the exclusion limit (1.0 ml) of this column. The dramatic change for 2,4-dinitronaphthol in Figure 3.9A indicates that the volume depended strongly upon the ratio of eluent component^.^ In acidic eluents (Figure 3.9C and D), the retention of acidic compounds becomes stronger and that of basic compounds becomes weaker. In this system, uric acid (h) is not suitable as a void volume marker due to its longer retention time. In neutral and basic eluents, an ionized acid can be used as the marker because no other compounds are eluted more rapidly. Fructose (c) is a very

Preparation, Testing, and Selectivity o Stationary Phase Materials f

45

mL

0.5 1 0

50

J 0.5 100 0

50

100

2.0

gi/
1.5
mL

2.5

1.o

1.oi

io

11.0 100 50 % conc. water in acetonitrile

100

Figure 3.9 Elution volumes ( m l ) of alternative void volume markers. Column, 5 pm octadecyl-bonded silica gel, 15 cm x 4.5 mm i.d.; eluents A and B, 10-90% aqueous acetonitrile, eluents C and D, 10-90% aqueous acetonitrile containing 50 mhfphosphoric acid;flow rate, 1 ml min- ; temperature, 30 C; detection, UV 210 nm and refractometer. Sample: a, acetonitrile; b, methanol; c, fructose; d, 2,4-dinitronaphthol; e, sodium nitrate; f, tetrahydrofuran; g, deuterium oxide; and h, uric acid.

polar compound, and is not adsorbed on a hydrophobic stationary phase surface. It is also UV-active at short wavelengths and non-ionizable under practical conditions. Its elution volume is nearly constant in the above conditions. The effect of injection volume and the necessity of a refractive index detector are thus avoidable. Therefore, fructose can be used as the standard void volume measuring marker. When analytes such as steroids are examined, they are excluded from many of the pores and the molecular size of fructose is too small to measure their void volume. Maltotriose may then be the compound of choice for the void volume measurement.

46

Chapter 3

6 Selectivity of Stationary Phase Materials

Silica Gel-based Stationary Phase Materials
Separations can be performed on a single column by using different eluents. Practical separations can also be achieved using column selection. In over 50% of applications octadecyl-bonded silica gel is used as the hydrophobic stationary phase material. As an alternative, phenyl-bonded silica gel is less hydrophobic compared with octadecyl- or octyl-bonded silica gel. This means that a smaller amount of organic modifier is necessary to obtain the same retentions. In addition, aromatic compounds are relatively more highly retained on phenylbonded silica gel than on octadecyl-bonded silica gel, and alkanes are more retained on octadecyl-bonded silica gel than on phenyl-bonded silica gel as shown in Figure 3.10. Such selectivity can be useful for the separation of compounds having different degrees of aromatici ty. Many manufacturers sell the same types of stationary phase materials,6 but the most popular stationary phase materials, e.g. octadecyl-bonded silica gels, from different sources, even from the same manufacturer, often demonstrate different retention capacities and selectivities. Such differences are due to the aggressive reactivity of the silylation method and the different bonding reactions that are used, as described earlier. The selectivity differences can be understood by the comparison of four octadecyl-bonded silica (ODS) phases synthesized from the same silica gel. The

O: + !; -0.5 -0.5
0

,
0.5

,
1

1
1.5

logk on octadecyl-bonded silica gels in 70% aq. acetonitrile

Figure 3.10 Selectivity of phenyl-bonded silica gel compared with ODs silica gel. Columns: ODs-silica, Y M C O D s , 15 cm x 6.0 mm id., phenyl-bonded silica gel, Y M C phenyl, 10 cm x 6.0 mm i.d.; eluent, aqueous acetonitrile. Compounds: x , polycyclic aromatic hydrocarbons; 0, alkylhenzenes; 0 ,polychlorobenzenes; + , alkanols; A, alkanes; 1, benzene; 2, benzopyrene; 3 , toluene; 4, heptylbenzene; 5 , hexachlorobenzene; 6 , hexanol; 7 , tetradecanol; 8, pentane; and 9, octane.

Preparation, Testing, and Selectivity of Stationary Phase Materials

47

four materials are a high-carbon (H1C)-loaded silica gel with and without a second silylation (end-capping) (HIC-ODs-E and HIC-ODs-NE, respectively) whose carbon content is 16 wt%, and a low-carbon (L0C)-loaded silica gel with and without a second silylation (LOC-ODs-E and LOC-ODs-NE, respectively) whose carbon content is 8.9 ~ t % The selectivity-related hydrophobicity . ~ differences can be compared using the differences in the retention factors of homologous alkylbenzenes as the van der Waals volume increases (Figure 3.1 1). The hydrophobicity, i.e. retention capacity, of HIC-ODS was found to be about twice that of LOC-ODs. The end-capping treatment further increased the retention capacity. This result indicates that if the quality control of the bonding processes is poor, the products will have different retention capacities. This difference also depends upon the surface area of silica gels. The relationship of the selectivity to the polarity of the analytes can be understood from the differences in the retention factors of homologous alkanols (Figure 3.12). The polar alkanols are relatively more retained on the nonendcapped bonded phases (LOC-ODs-NE and HIC-ODs-NE) because smaller-size alkanols can reach the unreacted silanol groups on the surface of silica gels. The relationship of the selectivity towards n-electrons can be understood from the differences in the retention factors of polycyclic aromatic hydrocarbons (Figure 3.13). The difference in the retention factors on end-capped and non-endcapped stationary phase materials is less than that of alkylbenzenes. This is due to the water content of the stationary phase. The content may be higher in non-endcapped bonded phases. The best octyl- and octadecyl-bonded silica gels should be chemically stable in highly basic solutions. The well-bonded phases give good results in the inertness

dl T-

'1
1 75 125 1 '5
cm3/mol

-1 25

van der Waals volume

Figure3.11

Comparison of the retention o toluene to decylbenzene as a measure o f f hydrophobicity of different octadecyl-bonded silica gels with different carbon LOC-ODs-E; 0 ,LOC-ODs-NE; 0, HIC-ODs-E; loadings. Columns: 0, A, HIC-ODs-NE; eluent, 80% aqueous acetonitrile at 30 "C. Compounds: 1, toluene; 2, nonylbenzene.

48

Chapter 3

8
rc

05 .

8
8 Q n n o
P O
0

0
0
0
2

0
U

-0.5

c ,

-1 I 50 75 100 125 1 i0

van der Waals volume cm3/mol

Figure 3.12 Polar selectivity of diflerent octadecyl-bonded silica gels towards the retention of homologous alkanols. Columns: 0, LOC-ODS-E; 0 , LOC-ODSNE; 0, HIC-ODs-E; A, HIC-ODs-NE; eluent, 70% aqueous acetonitrile at 30 "C. Compounds: 1, butanol; 2, dodecanol.

0
Y

8 c ,

0.5-

gA 0
0

rc

! i
E

O-

s
0
0 1

x o
0
2
I

0 .-

o o 4
3

c ,

s
2

-0.5-

+r .

-1

25

50

75

100

125

1 i0

van der Waals volume

cm3/mol

Figure 3.13 n-Electron selectivity of different octadecyl-bonded silica gels towards the retention of polycyclic aromatic hydrocarbons. Columns: 0, LOC-ODs-E; 0 , LOC-ODs-NE; 0, HIC-ODs-E, A,HIC-ODs-NE; eluent, 80% aqueous acetonitrile at 30 "C. Compounds: polycyclic aromatic hydrocarbons; 1, benzene; 2, naphthalene; 3 , pyrene; 4,3,4-benzopyrene.

tests (see above), and can be used for continuous analysis in basic solutions, pH 9.0, over a one month period. Such stability and inertness are guaranteed by the manufacturers. All silica-loaded materials are unstable in high pH eluents. The amount of carbon loading can be more than 110% of the theoretical calculation; however, the accuracy of surface area measurement is less than that of the carbon loading measurement. Non-endcapped bonded phases sometimes

Preparation, Testing, and Selectivity of Stationary Phase Materials

49

demonstrate better relative separations; however, their long-term operation is not guaranteed. The lifetime of short-chain bonded phases, such as butyl, cyano, and amino, is also short in buffer solutions due to their weak hydrophobicity. It seems that these phases do not reject the adsorption of ions from the eluent on to the untreated silanol groups.

Organic Polymer-based Stationary Phase Materials

Surface-modified silica gels are used for a variety of separations, but organic polymer-based stationary phase materials are more useful for long-term operations, such as for an amino acid analyser and size-exclusion liquid chromatography. Physically strong organic porous and non-porous polymer gels are also used for high-performance liquid chromatography. When difficulty in achieving separation is encountered with silica gel-based stationary phase materials, organic gels will often solve this problem even though the theoretical plate number may be less than that of silica gel-based stationary phase materials. For example, polystyrene gel is very hydrophobic, and aromatic compounds are selectively retained. Polystyrene oligomers are separated on a polystyrene gel using isocratic elution (Figure 3.14). Vitamins, anions, and nucleic acids and bases have been chromatographed on polystyrene gel.* These separations can also be carried out on alkyl group-bonded silica gels, if the silanol groups are completely ~ o v e r e dLarge pore-size polystyrene gels are useful for the separa.~ tion of proteins. Proteins with M , = 6000-66300 have been separated on a

min

35

Figure 3.14 Reversed-phase liquid chromatogram of polystxrene oligomers. Conditions: columns, PLRP-S, 15 cm x 4.6 mm id., 100 A (polystyrene gel); eluent, 70% aqueous tetrahydrofuran;Jow rate, 1 ml min-'; detection, UV 254 nm. Numbers beside peaks indicate the molecular weight. (Reproduced by permission from Polymer Laboratories data)

50

Chapter 3

300 pore size polystyrene gel, and M , = 120 000 collagen and 340000 fibrinogen have been separated on a 1000 pore-size polystyrene gel, as shown in Figure 3.15. Silica gel-based bonded stationary phase materials with a 1000 pore size are not as physically strong as the narrow pore-size silica gels. However, polystyrene gels have the advantages of surface inertness and moderate physical strength. Methacrylate polymers are as physically strong as polystyrene gels. Their hydrophobicity is weaker than that of polystyrene gel, and the aromatic selectivity is similar to that of octadecyl-bonded silica gel. A chromatogram of hydantoins on a methacrylate polymer is shown in Figure 3.16. The elution order is related to the hydrophobicity of the chemicals. Vinyl alcohol copolymer gel is hydrophilic and has been developed for aqueous-phase size-exclusion liquid chromatography; however, it is less polar than the polysaccharides. Its specificity permits the direct injection of a biological sample without deproteinization. For example, blood serum from a patient suffering from chronic nephritis has been injected directly as a measure of the degree of dialysis (Figure 3.17). Adenosine triphosphate, adenosine diphosphate, and adenosine monophosphate in red blood cells have also been separated directly (Figure 3.18). Theophylline in blood serum has been

1
I

J
I
12

4
I 0

I min 12

Figure 3.15 Chromatogram of jibre-type proteins polystyrene gels having diflerent pore sizes. Column A, PLRP-S 300 A, 15 cm x 4.6 mm i d . ; B, PLRP-S 1000 A (polystyrene gel), 15 cm x 4.6 mm i.d.; eluent, 15 min linear gradient from 20% of 0.25% triJuoroacetic acid to 60% of 0.25% trfluoroacetic acid in 95% aqueous acetonitrile; flow rate, 1.O ml min- detection, UV 220 nm. Peaks: 1, collagen ( M , 120 000) and 2,jibrinogen ( M , 340 000). (Reproduced by permission from Polymer Laboratories data)

';

Preparation, Testing, and Selectivity of Stationary Phase Materials

51

I 0

I 4

min

I 8

Figure 3.16 Chromatogram of hydantoins on polymethacrylate stationary phase. Conditions: columns, Shodex RSPak DE613 (methacrylate gel); eluent, 1:1 mixture o 0.033 M disodium hydrogen phosphate and potassium dihydrogen f phosphate; flow rate, 2.0 ml min-'; detection, U V 210 nm; temperature, 50C. Compounds: 1, R = (CH2)ZCOzH; 2, R = CH20H; 3, R = H; 4,R = CH,; and 5, R = C2H5. (Reproduced by permission from Shodex data)

1 2

I
0

L
20

min

I 40

Figure 3.17 Chromatogram of blood serum from a patient with chronic nephritis. Conditions: columns, Asahipak GS320 (vinyl alcohol copolymer gel), 50 cm x 7.6 mm i.d.; eluent, 0.1 M sodium phosphate containing 0.3 M sodium chloride pH 7.0; f l o w rate, 1 ml min-'; detection, UV 250 nm; direct injection of sample. Peaks: 1 , protein; 2, orotidine; 3, creatinine; and 4,uric acid. (Reproduced by permission from Asahipak data)

separated from proteins (Figure 3.19). In these separations, the proteins were eluted first and the targeted compounds eluted later; the quantitative analysis of the target compounds is thus easy. Pesticides in blood serum have also been directly analysed on vinyl alcohol copolymer gels, as shown in Figure 3.20.

52
1

Chapter 3

10 min 20

Figure 3.18 Adenosine phosphates in blood on vinyl polymer column. Conditions: column, Asahipak GS320 (vinyl alcohol copolymer gel), 50 cm x 7.6 mm i.d.; eluent, 0.1 M sodium phos hate bufler containing 3 M sodium chloride pH 7.O;flow rate, 1.O ml min- {detection, UV 260 nm. Peaks: 1, haemoglobin; 2, adenosine triphosphate; 3, adenosine diphosphate; and 4, adenosine mono-

phosphate. (Reproduced by permission from Asahipak data)

I 0
'igure 3.19

min

I 16

Theophylline in blood serum on vinyl alcohol polymer column. Conditions: column, Asahipak GS320 (vinyl alcohol copolymer gel), 50 cm x 7.6 mm i.d.; eluent, 0.01 M sodium acetate bufler pH 4.0 in 10% aqueous acetonitrile; j?ow rate, 2 ml min-'; detection, U V 280 nm. Peaks: 1, protein; 2, low M , impurity; and 3 , theophylline. (Reproduced by permission from Asahipak data)

These separations can be carried out using a silica-based bonded phase; however, the important advantage of organic polymer stationary phase materials is their chemical stability. The columns can be washed by using an alkaline solution after a certain number of injections. According to the chromatograms, the proteins in serum are completely eluted and nothing remains inside the column. However, the pressure drop in this type of analysis

Yreparation, Testing, and Selectivity of Stationary Phase Materials

'

53

i
0 min

30

Figure 3.20 Pesticide in rat blood serum on vinyl alcohol polymer column. Conditions: column, Asahipak GS320 (vinyl alcohol copolymer gel), 50 cm x 7.6 mm i.d.; eluent, 0.05 M phosphate bufer containing 0.2 M sodium chloride pH 7.2 in 20% aqueous methano1;JIow rate, 1.0 ml min-'; detection, UV 254 nm. Peaks: 1, blood serum protein and 2, methomyl. (Reproduced by permission from Asahipak data)

may increase after several injections, depending on the injection volume. This problem is not avoidable on all types of columns. Therefore, washing using an alkaline solution is occasionally required. The separation of basic and metal-sensitive compounds is difficult on silicabased stationary phase materials, but these separations can be performed on vinyl alcohol copolymer gels. Examples are the separation of methallothionein from dolphin kidney, a-, b-, and y-endorphin, and nucleotide and nucleoside mixtures.8 However, an analytical-scale separation may also be performed on surface-modified wide-pore silica gels (pore size 300 A or more), using columns which showed a negative response in the heavy metal test described above. Modified vinyl alcohol copolymer gels with phenyl groups improved the hydrophobicity and selectivity for aromatic compounds. Chromatograms of protein mixtures and enzymes (phosphoisomerase) are shown in Figures 3.2 1 and 3.22, respectively. The separation mechanism on the surface of these stationary phase materials is basically the same as that for reversed-phase liquid chromatography, and is called hydrophobic liquid chromatography. The chromatograms of protein mixtures shown in Figure 3.23 were obtained on a phenyl-bonded vinyl alcohol copolymer gel and a trimethylsilyl-bonded silica gel using basically the same eluent. The elution order of proteins was the same, but the resolution was a little different. Octadecyl-bonded vinyl alcohol copolymer gels are chemically stable and overcome the weakness of octadecylbonded silica gels, which are influenced by unreacted silanol groups and high pH instability.' The selection of inorganic or organic polymer stationary phase materials depends on the application. When information is needed rapidly, silica-based bonded phases have the advantage due to their physical strength and high theoretical plate number, although their lifetime is shorter. When a longer-term operation is necessary, such as for an amino acid analyser, organic polymerbased stationary phase materials are appropriate. Organic polymer-based

54

Chapter 3
1 4

I 0

30

min

I 60

Figure 3.21 Separation o proteins on wide pore phenyl-bonded column. Conditions: f column, TSKgel Phenyl-5 P W , 75 mm x 7.5 mm i.d.; eluent, 60 rnin linear gradient of ammonium sulfate concentration from 1.7 to 0 M in 0.1 M phosphate bufler (pH 7.0);JIow rate, 1.O ml min- detection, UV 280 nm. Peaks: 1, cytochrome c; 2, myoglobin; 3, ribonuclease; 4, lysozyme; 5, achymotrypsinogen; and 6, a-chymotrypsin. (Reproduced by permission from Tosoh data)

';

I 0

I 15

rnin

30

Figure 3.22 Chromatogram o phosphoisomerase. Conditions: column, TSKgel Phenyl-5 f P W RP, 75 mm x 4.6 mm i.d.; eluent, 2 min linear gradient o acetonitrile f from 0.05 to 20% followed by 48 min linear gradient of acetonitrile from 20 to 80% in 0.05% trljluoroacetic acid;_flow rate, 1 ml min-'; detection, UV 220 nm. (Reproduced by permission from Tosoh data)

stationary phase materials can be recycled, but silica-based stationary phase materials are used only once; this is another selection point, which depends on the economic considerations. The disposal of used stationary phase materials, especially those used for biomedical applications, should also be of concern. Silica gel-based stationary phase materials can be heated at high temperature before disposal, and organic polymer-based stationary phase materials can be burned. Take care.

Other Stationary Phase Materials

Hydroxyapatite and titania are also useful stationary phase materials for biological samples. Hydroxyapatite, Ca10(P04)6(OH)2, a hexagonal column as

Preparation, Testing, and Selectivity of Stationary Phase Materials

55

15

30

I min 15

30

f Figure 3.23 Selectivity o phenyl and alkyl bonded stationary phase materials for protein separation. Column A, TSK gel phenyl-5PW R P , 75 mm x 4.6 mm i.d.; B, T S K gel T M S 250, 75 mm x 4.6 mm i.d.; eluent, 60 min linear gradient elution from 5% of 0.05% trfluoroacetic acid in 5% aqueous acetonitrile to 80% o 0.05% trfluoroacetic acid in 80% aqueous acetonitrile; flow rate, f 1 ml min-'; detection, UV 220 nm. Peaks: 1, ribonuclease; 2, insulin; 3, cytochrome c; 4, lysozyme; 5, transferrin; 6, bovine serum albumin; 7, myoglobin; and 8, ovalbumin. (Reproduced by permission from Tosoh data)

1 200

E
E

I 0

I 10 min

L ! O I 20

s 5 m

Figure 3.24 Separation of plasmid D N A on a titania column. Conditions: column, H C A column A-7610 (titania), 10 cm x 7.6 mm i.d.; eluent, step-wise gradient o potassium phosphate bufer concentration (pH 6.8);flowrate, 1.O ml min- I!, detection, UV 260 nm. Peaks: 1, purasmid D N A (pBR 322) and 2, cromorimal. (Reproduced by permission from Polymer Laboratories data)

56

Chapter 3

crystal is formed as a porous stationary phase material. The large pore size, 70010 000 is suitable for purification of bio-polymers. Bovine serum albumin, lactic dehydrogenase, ribonucleic acid, and immunoglobulin in blood serum can be purified. Hydroxyapatite is also used as a support for affinity liquid chromatography . The pore size of porous titania can be up to 2000 Titania is used for the purification of proteins and as a support for bound enzymes. The purification of P-lactoglobulin from cheese whey, of protease from pineapple, /.I-lactamase, and amylase can be achieved with titania. The latter two purifications are impossible on alumina. Titania is also used as a support in peptide synthesis. The separation of plasmid DNA is shown in Figure 3.24.

A,

A.

7 References
1. G.B. Alexander (ed.), Silica and I,, Doubleday & Company, 1967. 2. K.K. Unger (ed.), Porous Silica, Journal of Chromatography Library, Vol. 16, Elsevier, Amsterdam, 1979. 3. M. Ohhira, F. Ohmura, and T. Hanai, J . Liq. Chromatogr.,l989, 12, 1065. 4. T. Hanai, M. Ohhira, and T. Tamura, LC-GC, 1988,6,922. 5. L. Boksanyi, 0. Liardon, and E. sz Kovats, Adv. Colloid Interface Sci., 1976, 6, 95. 6. T. Hanai, in Encyclopedia of Analytical Science, Academic Press, London, 1995, pp. 2558-2567. 7. J. Yamaguchi, T. Hanai, and Hong Cai, J . Chromatogr., 1988,441, 183. 8. Asahipak data from Showa Denko. 9. K. Yasukawa, Y. Tamura, T. Uchida, Y. Yanagihara, and K. Noguchi, J. Chromatogr., 1987,410, 129.

CHAPTER 4

Selection of the Eluent

The most important parameter for the control of liquid chromatography is the composition of the eluent. Liquid chromatography is a powerful separation method with unlimited possibilities of eluent selection. However, it is not easy to choose a suitable eluent within a short time without a number of trial experiments. The crucial factor is to control the solubility of the analytes in the eluent. Increasing the solubility of analytes in the eluent decreases their retention times. The selection of the components of an eluent is described below, based on the properties of the analytes to be separated. The important properties are hydrophobicity, dipole moment, hydrogen bonding, ionization, and steric effects.

1 Reversed-phase Liquid Chromatography

The most popular system, with over 50% of separations, is reversed-phase liquid chromatography on octadecyl-bonded silica gel. Analytes are retained on the stationary phase materials based on their hydrophobicity. This means that polar compounds are eluted faster than non-polar compounds. The higher the hydrophobicity of the stationary phase surface, the longer will be the retention times of the analytes. Bonding of donor-electron groups onto the stationary phase surface results in a stronger retention of compounds containing dipoles. In some cases, interactions between untreated silanol groups on the stationary phase and hydroxy groups on the analyte increase the selectivity; however, polar groups on the stationary phase are generally eliminated to obtain inert stationary phase materials, which are more stable in long-term operation.

Selectivity of Organic Modifiers in the Eluent

Three organic solvents, acetonitrile, methanol, and THF, are usually used as the organic modifiers. Increasing the concentration of the organic modifier decreases the overall retention times, but changes in relative retention times depend on the properties of the analytes.
57

58

Chapter 4

The general relationship between the type of solute and its retention can be seen by comparing the retention factors, k,of a set of standard compounds with their octanol-water partition coefficients, i.e. the log P value (listed in Table 4.1), as a measure of their relative solubility in water. The logarithm of the retention factor, logk, of these compounds measured in 50% aqueous acetonitrile on an octadecyl-bonded silica gel column shows a close linear relationship (Figure 4.1). However, the elution behaviour of aromatic and aliphatic compounds is often different (even though their carbon numbers and van der Waals volumes are very similar) in eluents containing different organic modifiers. These variations are due to differences in the solubility of analytes in the organic solvent. Five organic solvents [acet onit rile, methanol , tetrah ydro furan (THF), acetone, and dimethylformamide], which are homogeneously miscible with water, have been used as modifiers to study the relationship of the selectivity of the solvent to the molecular properties of analytes. The polar interaction Table 4.1 Log P values of standard analytes
~

No. Chemical
1 2 3 4 5 6 7 8 9 10

Log P
1.33 1.86 2.39 2.92 3.98 5.04 2.15 3.09 3.74 4.14

No. Chemical
11 12 13 14 15 16 17 18 19

Log P
2.28 3.21 3.91 4.38 5.03 5.79 2.59 3.12 3.52

Pentanol Hexanol Heptanol Octanol Decanol Dodecanol Methyl benzoate Isopropyl benzoate Butyl benzoate Isopentyl benzoate

Benzene Naphthalene Fluorene An thracene Pyrene Chrysene Toluene Ethy1benzene Isopropyl benzene

o!
1

,
2

log P
Figure 4.1 Retention factors related to log P values. Column, 5pm octadecyl-bonded silica gel (LiChrosorb LC7) 25 cm x 4.1 mm i.d.; eluent, 50% aqueous acetoalkanols; 0 , benzoates; nitrile; j l o w rate, 1 ml min-. Compounds: 0, 0, polycyclic aromatic hydrocarbons; and A, alkylbenzenes.

Selection of the Eluent

59

Table 4.2 Interaction properties of solvents for reversedphase liquid chromatography

Solvent Tetrahydrofuran Acetone Methanol Acetonitrile Dimethylformamide Water

Po
4.0 5.1

Xe
1.52 1.78 2.45 1.80 2.50 3.77

Xd 0.80
1.17 1.12 1.57 1.34 3.77

Xn
1.68 2.14 1.58 2.44 2.56 2.55

5.1
5.8 6.4 10.2

Po: polarity, Xe: proton acceptor, Xd: proton donor, Xn; dipole moment.

properties of the solvents were calculated from reference values,' and are listed in Table 4.2. The values of the polarity (Po)and the polar interaction properties, proton acceptor ( X e ) , proton donor ( X d ) , and dipole moment (Xn), of the mixed solvents can be calculated from their concentrations. As above, the logk values of the different types of compounds [alkanols (ROH), polycyclic aromatic hydrocarbons (PAH), alkyl benzenes (RB), and alkyl benzoates (ROB)] each demonstrated a linear relationship with their log P values and this relationship was observed with different ratios of acetonitrile and water. The experiment was then performed in eluents containing different organic modifiers and it was found that the behaviour of analytes containing different functional groups differed and the effect depended on the l o g P of the analy t e. To demonstrate these differences, the experimental relationship between log P and log k across a range of eluent compositions was determined for each group of analytes. The results were then used to calculate the predicted logk in each case for a theoretical model compound with log P = 5 as a hydrophobic model, for l o g P = 1 as a hydrophilic model, and for logP = 3 as an intermediate model. The results in a range of aqueous acetonitrile eluents are shown in Figure 4.2. For both alkanols and polycyclic aromatic hydrocarbons the predicted values for each of the model compounds fall on a single line represented by the solid points ( 0 ) (as expected from Figure 4.1). However, with T H F as the modifier, predicted values for the model alkanols (ROH) (+) and polycyclic aromatic differ hydrocarbons (PAH) (0) and the relative positions change with the log P values. The log k values of log P = 5 alcohols and PAHs in the aqueous T H F eluent were relatively smaller than those obtained for the alkanols and PAHs in aqueous 30-70% acetonitrile, but the logk values of the l o g P = 1 model compounds were higher. Furthermore, in aqueous T H F the logk values of polycyclic hydrocarbons were larger than those of alkanols. These results indicate that when a separation between aromatic and polar aliphatic compounds is necessary, T H F may be an effective organic modifier.

60
logPr5
2.0

Chapter 4
log P = 3 log P = 1

8 - 1.0

0 0

50

10WO

50

10010

50

100

concentration of organic modifier / water Figure 4.2

Predicted chromatographic behaviour in aqueous acetonitrile and tetrahydrofuran o model compounds whose log P values are 1, 3, and 5. Column, 5 pm f octadecyl-bonded silica gel (LiChrosorb LC-7), 25 cm x 4.1 mm i d . Compounds and eluent: a, polycyclic aromatic hydrocarbons and alkanols in aaueous acetonitrile: 0 alkanols in aaueous -tPtrahvdrnfuran and n) r - - j -- - -- -, " __ --- - - __ - - - . __ . , -~-_. -, --- v nnlv, -. cyclic aromatic hydrocarbons in aqueous tetrahydrofuran.
-. -. . . . . . .... ..
'

Comparison of the Efect o Organic ModiJiers on Analytes with f logP = 3

The selectivity of a number of organic modifiers was examined using the predicted log k values of the log P = 3 models from each group in different organic modifier-water mixtures. The composition of the eluent was adjusted so that either the solubility parameter,' polarity (Po),proton acceptor ( X a ) ,proton donor ( X d ) , or dipole moment ( X n ) values were kept constant to determine which parameter affected the selectivity. The results are summarized in Table 4.3. The compositions of the different eluents were determined using the values in Table 4.2. The composition was adjusted in each case so that the selected polarity term was equal in each eluent. For example, in Table 4.3 the polarity (Po) of the 40 :60 THF-H20 is (0.4 x 4) + (0.6 x 10.2) = 7.72, when X e , 2.87; X d , 2.58; Xn, 2.20. This has an equivalent polarity to 56:44 ACN-H20 [(0.56 x 5.8) + (0.44 x 10.2) = 7.72 but now X e , 2.66; Xd, 2.53; X n , 2.491. Chromatographic Behaviour of log P = 3 Compounds in Aqueous Tetrahydrofuran. For analytes whose log P = 3, in 40% aqueous T H F the predicted log k values of the model alkanols (ROH), polycyclic hydrocarbons (PAH), alkylbenzenes (RB), and alkyl benzoates (ROB) are 0.64, 0.77, 0.90, and 0.79, respectively. The polarity of 40% aqueous T H F is 7.72. The same polarity can be obtained in 56% aqueous acetonitrile. As noted earlier, in aqueous aceto-

Selection of the Eluent

61

Table 4.3 Solubility problems of eluents

Predicted log k of log P compound Eluent Po Xe Xd Xn ROH 0.64 PAH 0.77 RB 0.90 0.40 0.63 0.46 nc 36 0.67 0.18 0.78 0.14 nc 44 0.41 0.37 0.53 0.36 56 0.87 0.70 1.02 0.43 nc 37
=

ROB 0.79

40% THF-H20 7.72 2.87 2.58 2.20 7.72 2.66 2.53 2.49 56% ACN-H20 8.19 2.87 2.76 2.50 46% ACN-H20 54% ACN-H20 7.82 2.70 2.58 2.49 nc nc nc nc 2.20 YOof ACN-H20 to give the same log k as 40% THF

46

40 0.63

43 0.62

60% MeOH-H20 7.14 2.98 2.18 1.97 0.81 7.14 2.40 2.24 2.47 69% ACN-H20 8.44 2.98 2.89 2.50 40% ACN-H20 7.02 2.35 2.18 2.47 72% ACN-HO nc nc nc nc 1.97 YOof ACN-H20 to give the same log k as 60% MeOH 39
7.65 2.78 2.47 2.35 50% ATN-HzO 58% ACN-H20 7.65 2.63 2.50 2.49 50% ACN-H20 7.99 2.78 2.66 2.49 59% ACN-H2O 7.60 2.61 2.47 2.48 "/o of ACN-H20 to give the same log k as 50% ATN

46 0.33

46 0.31

0.35

59

61 0.69

62 0.89

50% DMF-H20 8.30 3.13 2.56 2.55 1.26 43% ACN-H20 8.30 2.92 2.82 2.50 8.77 3.13 3.06 2.51 33% ACN-H20 7.78 2.69 2.56 2.49 55% ACN-H20 10.2 3.77 3.77 2.55 0% ACN-H20 YOof ACN-H20 to give the same log k as 50% DMF 26

44

36

Po: Polarity, Xe: proton acceptor, Xd: proton donor, Xn: dipole moment, THF: tetrahydrofuran, ACN: acetonitrile, MeOH: methanol, ATN: acetone, DMF: dimethylformamide, nc: cannot be calculated.

nitrile the log k values for all compounds with log P = 3 are the same and thus the predicted value for the ROH, PAH, RB, and ROB models is 0.40, unlike any of the T H F retentions. These results indicate that chromatographic behaviour cannot be controlled only by polarity. The proton acceptor value (Xe) of the 40% aqueous T H F eluent was 2.87. The same proton acceptor value can be obtained in 46% aqueous acetonitrile (Table 4.3), where the predicted logk values of each of the log P = 3 model compounds is 0.63. Thus in this case the value of the logk of ROH is nearly equal to the value in 40% aqueous T H F (0.64). If the proton donor value (Xd) fixed at 2.58, the equivalent eluent is 54% is aqueous acetonitrile and the predicted log k value for the log P = 3 compounds should be 0.46, which is again different from any of the logk values in 40% aqueous THF. The dipole moment ( X n ) of 40% aqueous THF was 2.20; however, such a value cannot be obtained with any proportion of aqueous acetonitrile. To obtain the same logk values as in 40% aqueous THF, a

62

Chapter 4

composition of 46% aqueous acetonitrile is needed for ROH, 40% aqueous acetonitrile for PAH, 36% aqueous acetonitrile for RB and 43% aqueous acetonitrile for ROB. This comparison demonstrates the different selectivity of THF.
Chromatographic Behaviour of log P = 3 Compounds in Aqueous Methanol. In a similar calculation, the logk values of ROH, PAH, RB, and ROB in 60% aqueous methanol are 0.81, 0.63, 0.67, and 0.62, respectively (Table 4.3). The same log k values can be obtained in 39% aqueous acetonitrile for ROH, in 46% for PAH, 44% for RB and in 46% for ROB. The values of Po, Xe, and Xd (but not Xn) in 60% aqueous methanol can be obtained in 69% for Po, in 40% for Xe, and in 72% aqueous acetonitrile for Xd. Only the logk of ROH in 40% aqueous acetonitrile demonstrated a nearly equal log k value to that obtained in 60% aqueous methanol. Thus the solvent properties do not have a specific relationship with the log k values of these compounds. Chromatographic Behaviour of log P = 3 Compounds in Aqueous Acetone. When the Xe value in aqueous methanol was nearly equal to that in aqueous acetonitrile (see above), the logk of ROH was about the same as that in aqueous acetonitrile but the corresponding predicted value with aqueous acetone (log k = 0.35) is very different to the equivalent aqueous acetonitrile (logk = 0.53) (Table 4.3). However, when the Po and X d values in aqueous acetonitrile are nearly equal to those in aqueous acetone, the predicted logk values of OH are about the same. To obtain the same logk as in 50% aqueous acetone it was necessary to use 59% aqueous acetonitrile for ROH, 61 YOfor PAH, 56% for RB and 62% for ROB. Chromatographic Behaviour of log P = 3 Compounds in Aqueous Dimethylformamide (DMF). DMF is a solvent with a strong dipole moment (Table 4.2). The same logk values in 50% aqueous DMF can be obtained at a concentration of 26% aqueous acetonitrile for ROH, 44% aqueous acetonitrile for PAH, 37% for RB, and 36% aqueous acetonitrile for ROB. When eluents with equivalent Po, Xe, Xd, or X n values in 50% DMF were selected, different log k values were predicted in all three cases. The above results only demonstrate the chromatographic behaviour of compounds with log P = 3. Their relative retention factors thus are influenced by the properties of solvents in the eluent. The relative values cannot be controlled by only one property of the solvents and chromatographic behaviour therefore also depends on the properties of the analytes.

Comparison of Effects of Mod$ers on a Mixture ofAnalytes

The following results demonstrate the comparative chromatographic behaviour of the 19 compounds listed in Table 4.2. In reversed-phase liquid chromatography, changing the concentration of organic modifier affects both the overall retention time and the relative elution behaviour of different types of com-

Selection of the Eluent

63

pounds. The effect of the organic modifier especially influences the retention of polycyclic aromatic hydrocarbons, as seen in Figure 4.3A (compounds 11-16). The results in 50% aqueous acetone and aqueous acetonitrile indicate that the retention times in 50% aqueous acetone are shorter than those in 50% aqueous acetonitrile. A selectivity difference for alkanols ( 0 , 1-6) is not observed but the retention time of the larger PAHs is shorter than the values expected from their size. From the results in 40% aqueous T H F and 50% aqueous acetonitrile (Figure 4.3B), the relative retention of PAHs (1 1-16) is weak and that of alkylbenzenes (0, 17-19) is strong in aqueous THF. The alkyl group affects the strong retention in aqueous THF. As found earlier (Table 4.3), a methanol-water

/ "12 9 13

-7-

11

0.5

1.0

1.5

L
0

log k 50% acetonitrildwater

0.5 1.0 1.5 log k 50% acetonitrilehater

C
1.5

0 0

0.5

1.0

1.5

0.5

1.0

1.5

log k 50% acetonitrildwater

log k 50% acetonitrildwater

Figure 4.3 Relative selectivity of diflerent organic mod$ers. Numbers beside symbols, 1-6 ( 0 ) ROH; 7-10 ( 0 ) ROB; 11-16 (filled hexagons) PAH; and 17-19 (0) RB.

64

Chapter 4

mixture (Figure 4.3C) can distinguish alkanols (e, ROH) from the other analyte groups. Comparison of the retention behaviour in 50% aqueous D M F and acetonitrile (Figure 4.3D) indicated that DMF is a selective modifier for the separation of alkanols ( 0 , 1-6) and PAHs (11-16) and of aromatic and aliphatic compounds. Unfortunately, DMF is too aggressive as a solvent and attacks organic polymer piston seals in the pump and, therefore, highly concentrated solutions cannot be used as an eluent. From the theoretical viewpoint, acetonitrile is the most suitable solvent to study the correlation of retention times and log P values of analytes, since the dipole moment (2.44) is nearly equal to that of water (2.55) (Figure 4.4). The electron donor effect can therefore be eliminated, and the elution order is not changed on modification of the acetonitrile-water mixture ratio. The first choice of an eluent should therefore be an acetonitrile-water mixture for non-ionic compounds in reversed-phase liquid chromatography. Methanol, acetone, THF, or DMF can then be added to improve the resolution.
Example I : Separation of an alkylbenzene and an alkyl benzoate. The logP values of butyl benzoate and isopropylbenzene are 3.74 and 3.52, respectively, from Table 4.1. The separation of these compounds in 50% aqueous acetonitrile therefore requires a high plate number column as the log k values will be nearly identical. The most suitable eluent is 40% aqueous THF. See compounds 9 and 19 in Figure 4.3B. Example 2: Separation of a polycyclic aromatic hydrocarbon and an alkyl benzoate. The log P values of fluorene and butyl benzoate are 3.91 and 3.74, respectively, from Table 4.1. The separation is poor in 50% aqueous aceto-

10

50

100

conc. Water / organic solvent

Figure 4.4 Polarity, proton acceptor and dipole properties of water and organic modifier mixtures. , acetonitrile; - - - , methanol; and . tetrahydrofuran.
*

Selection of the Eluent

65

nitrile; however, they can be separated in 60% aqueous methanol or 50% aqueous DMF (compounds 13 and 9, Figures 4.3C and 4.3D). However, in selecting a mobile phase the entire analytical system and separation system should be considered. In many cases, a component of a good eluent may not be suitable for research purposes. For example, the use of water as an eluent in methods that will require further chemical reactions and purification should be avoided due to the difficulty in concentrating the products of interest after the separation and the stability of the products in the process. In addition, organic solvents are flammable, combustible, and toxic. The threshold limit concentrations of acetone, acetonitrile, DMF, methanol, and T H F are 250, 20, 10, 200, and 200ppm, respectively. The most easily handled solvent of these five is methanol. Acetone is the least used solvent due to the strong UV absorption (UV cut-off = 330 nm). T H F is unstable and easily oxidized by oxygen in air, causing a cut-off at longer wavelengths as it ages. At present, the selection of an organic modifier is estimated from the aliphatic or aromatic nature of analytes. However, the properties of analytes often cannot be easily obtained. Examples of quantitative structure-retention relationships based on the lo g P and van der Waals volume of analytes are demonstrated in Chapter 6.

2 Chromatography of Ionic Compounds

Neutral compounds are retained by hydrophobicity in reversed-phase liquid chromatography, and by hydrogen bonding and/or n-n interactions on hydrophilic phases in normal-phase liquid chromatography. Ionization of the analyte reduces the retention time in reversed-phase liquid chromatography, but in some cases increases the retention time on ion-exchangers. Retentions on ionexchange resins made from polystyrene gel are based both on ion-ion interactions with ion-exchange groups and on hydrophobic interactions with the polystyrene gel matrix. The addition of counter-ions can improve the resolution.

pH Effects
The retention factor of partially ionized compounds can be predicted in reversed-phase liquid chromatography by Equation (4. 1):2

where ko is the maximum retention factor of the unionized form of the analyte. The term ki is the retention factor of the fully ionized compound; K, is the dissociation ionization constant; and [H '1 is the hydrogen ion concentration of the eluent. Every compound has its own ko, ki, and K, values. When the hydrophobicities of compounds are nearly equal, separation is difficult in the

66

Chapter 4

reversed-phase mode. However, when their dissociation constants are different, the separation can be easily accomplished in a pH-controlled eluent by introducing differential partial ionization. The value of ki is close to zero in a reversed-phase system; however, the value is not negligible in the ion-exchange mode.

Example: separation of 3-chloro and 4-ethylbenzoic acids. The log P values of these compounds are similar, and their pKa values are 3.96 and 4.35, respectively. As seen by their chromatographic behaviour in Figure 4.5, the separation is difficult at low pH due to the similar log P values giving similar log ko values. The separation can be easily performed in an eluent with a pH of about 4.2, due to the difference in PKa values and different degrees of ionization. The retention factor of an amino acid is given by the following equation, as the amino acid has two ionizable groups (a carboxyl group and an amino group, with Kal and Ka2,respectively):

The maximum retention factor (ko) is related to the log P value and kl and k- 1 are the retention factors of the cationic and anionic forms, respectively. The pKa values are known, and the retention factor in a given eluent can therefore be predicted in reversed-phase liquid chromatography using an alkyl-bonded silica gel or polystyrene gel column. The separation conditions can be adjusted according to their logP and pKa values by the selection of a suitable organic modifier concentration and the pH of the eluent.374

I 2
Figure 4.5

I 4 PH

I 6

Eflect of p H on the retention factors of 3-chloro- and 4-ethylbenzoic acids. Column, octadecyl-bonded vinyl alcohol copolymer gel, 10 cm x 6 mm i d . ; eluent, 0.05 M sodium phosphate solution in 20% acetonitrile; flow rate: 1 ml min- '; temperature, 30 "C. Compounds: 3C1, 3-chiorobenzoic acid; and 4Et, 4-ethylbenzoic acid.

Selection of the Eluent

67

The separation of compounds whose pK, values are different can be carried out easily on an ion-exchanger using ion-ion interaction^.^ For example, nonionized mandelic acid is eluted before benzoic acid due to weak hydrophobicity. But mandelic acid is retained more than benzoic acid on an ion-exchanger due to its stronger acidity. With polystyrene gel-based ion-exchangers both ion-ion and hydrophobic interactions are important factors in the retention mechanism. Compounds whose pK, values are similar can be separated on an ion-exchange resin. For example, amino acids have been separated by cation-exchange liquid chromatography (Figure 4.6) and carboxylic acids by anion-exchange liquid chromatography (Figure 4.7). The elution order of these separations was directly related to their dissociation constants. The high temperature of the separation was also important, as were the acidic and ionic strength of the eluen ts. The separation of very hydrophobic compounds, such as proteins, is difficult on an ion-exchange resin with a hydrophobic matrix, because organic modifiers are required as part of the eluent. However, the separation can be performed on an ion-exchanger with a hydrophilic matrix. A weak acidic ion-exchanger is suitable for the separation of basic proteins, and a weak basic ion-exchanger is suitable for the separation of acidic proteins (Figures 4.8 and 4.9, respectively). The components of the eluents should not be harsh, to prevent the denaturation of the proteins. A salting-out effect can be used to obtain a gradient elution. A combination of buffer solutions and weak ion-exchangers on a polar matrix, such as poly(viny1 alcohol), is suitable for the separation of bio-polymers, especially for preparative-scale separations. Oligonucleotides from polyadenyl acid and RNA of Escherichia coli have been separated on weak anionexchangers using ammonium acetate buffer and a sodium chloride concent ra1 2 12 13 1718202123 25

I
0

I 20

I 40

I
min 60

I 80

Figure 4.6

Separation of amino acids in body fluid. Column, M C I G E L CKIOV, 15 cm x 4.6 mm i.d.; eluent, four-step gradient o sodium citrate bufler;flow f rate, 0.5 ml min- '; temperature, 60 "C; detection, fluorescence reaction detector of sodium hydrochloritelorthophthaldehyde, Ex. 360 nm, Em. 450 nm. Peaks: 1, S S A ; 2, T A U ; 3, GCGA; 4,HPro; 5 , Asp; 6 , Gln + Asp; 7 , Thr; 8, Ser; 9, Sar; 10, Glu; 11, Pro; 12, Gly; 13, Ala; 14, Cys; 15, Val; 16, Met; 17, Ile; 18, Leu; 19, Tyr; 20, Phe; 21, His; 22, Orn; 23, Lys; 24, Trp; and 25, Arg. (Reproduced by permission from Mitsubishikasei data)

68
2

Chapter 4

I
0

I
10

I I 20 min 30

Figure 4.7 Anion exchange separation o carboxylic acids in red wine. Column, Shodex f C811, 100 cm x 7.6 mm id.; eluent, 3 mM perchloric acid; flow rate, 0.9 ml min- temperature, 60 "C; detection, reaction detection using chlorophenol red at 430 nm. Peaks: 1, citric acid; 2, tartaric acid; 3, malic acid; 4, succinic acid; 5, lactic acid; 6, formic acid; and 7 , acetic acid. (Reproduced by permission from Shodex data)

';

lj
I
L

I 0
Figure 4.8

I I 20 rnin 40

Cation-exchange liquid chromatography o basic proteins. Column, Asahipak f ES502C; eluent, 20 min linear gradient o sodium chloride f r o m 0 to 500 m f M in 50 mM sodium phosphate bufer pH 7.0;JEow rate, 1 ml min-'; temperacure, 30 "C; detection, UV 280 nm. Peaks: l ? myoglohin f r o m horse skeletal muscle ( M , 17 500, PI 6.8-7.3); 2, ribonuclease f r o m bovine pancreas ( M , 13700, PI 9.5-9.6); 3, a-chymotrypsinogen A f r o m bovine pancreas ( M , 257 000, PI 9.5); and 4,lysozyme from egg white (Mr 14 300, PI 11.O-11.4). (Reproduced by permission from Asahikasei data)

Selection of the Eluent

2

69

I
0

I
20 min

I 40

f Figure 4.9 Anion-exchange liquid chromatography o acidic proteins. Column, Asahipak ES502N; eluent, 20 min linear gradient of sodium chloridefrom 0 to 500 PM in 50 mM bis-trisfHC1 bufer pH 7.0; flow rate, 1 ml min-'; detection, UV 280 nm. Peaks: 1, conalbumin ( M , 7700&88000, PI 6.0-6.8); 2, ovalbumin ( M , 45000, PI 4.6); 3, trypsin inhibitor ( M , 8000, PI 4.5); and 4, plactoglobulin (Mr 18 400, pl 5.1). (Reproduced by permission from Asahikasei data)

tion gradient.697 Subfragments of myosin have been separated on a weak cation exchanger using a sodium chloride concentration gradient.* The selection of eluent components such as pH and salt concentration makes possible the direct analysis of catecholamines in urine without deproteinization and of haemoglobin in red blood by cation-exchange liquid chromatography, as shown in Figures 4.10 and 4.1 1, respectively.

c
I

min

50

Figure 4.10 Direct analysis of catecholamines in urine sample. Column, Asahipak ES502C; eluent, 75 m succinic acid + 25 mM borate bufer (pH 6.10) containM ing 0.5 r n M EDTA; flow rate, 1.0 min-'; detection, fluorescence reaction detection Ex. 350 nm. Peaks: 1, adrenaline; 2, noradrenaline; and 3, dopamine. (Reproduced by permission from Asahikasei data)

70

Chapter 4

I 0

min

I0

,; i
I

min

30

Figure 4.11 Direct injection analysis o haemoglobin in red blood cells. Column, Asahipak f ES-502C; eluent, 32 min linear gradient from 25% 30 mM sodium phosphate bufler to 65% 30 mM sodium phosphate containing 300 m M sodium chloride pH 5.5; $ow rate, 1.0 ml min-; detection, 425 nm. Samples: A, normal subject and B, diabetic patient.

Ion-pair Separations
Ionic compounds are usually separated by ion-exchange liquid chromatography, and non-ionic compounds are separated by normal-phase or reversedphase liquid chromatography. The separations of mixtures of ionic and nonionic compounds and of different ionic and polar compounds are more difficult. One method for extracting such compounds from aqueous solution is ion-pair extraction where the ionic charge on the analyte is neutralized by the addition of a counter-ion of opposite charge, and the analyte is extracted into an organic phase as the neutral ion-pair. This ion-pair extraction method was first applied as ion-pair partition liquid chromatography by SchilI et a1. The selection of the counter-ion and its concentration are important for the separation of ionic compounds in reversed-phase and ion-exchange liquid chromatography. The addition of hydrophobic ions is an especially powerful method and several surfactants can be used as hydrophobic counter-ions. The theoretical column efficiency of ion-pair liquid chromatography is much better than that of an ion-exchange column, and the regeneration of a column is much faster. Thus, if we can control ion-pair liquid chromatography, we can solve a separation problem. (The important background sources in this area are listed at the end of the chapter.)

Theory of Ion-pair Liquid Chromatography

An ionized solute A - forms a paired ion AB with a counter-ion B + in water. The paired ion is extracted into an organic phase from the aqueous phase, as given by the equation:

Selection of the Eluent

71

where aq is the aqueous phase and org is the organic phase. When applying this ion-pair extraction, the following parameters must be considered. a. b. c. d. e. f. Selection of stationary phase material. Selection of counter-ion and the concentration. pH and ionic strength of eluent. Selection of organic modifier and the concentration. Selection of column temperature. Selection of detector.

Two types of system are used for ion-pair liquid chromatography. When polar stationary phase materials, such as silica gel, are used an ion-pair partition mechanism is applied. When non-polar stationary phase materials, such as octadecyl-bonded silica gel and polystyrene gel, are employed a paired-ion adsorption mechanism is involved. The former is called normal-phase ion-pair partition liquid chromatography, and the latter is called reversed-phase ion-pair liquid chromatography. The following conditions shoudbe satisfied to perform satisfactory ion-pair liquid chromatography. a. A one-to-one ion-pair must be formed. b. The equilibrium concentration of the ion-pair in the aqueous phase must be very low compared with that of the ion. c. Only the paired-ion is extracted into the organic phase. d. The counter-ion makes an ion-pair only with the target ion. e. The paired-ion is not decomposed in the eluent. f. The partition equilibrium is related only to the chromatographic retention.

Normal-phase Ion-pair Partition Liquid Chromatography

Analytes in the organic phase (eluent) form ion-pairs with counter-ions existing in the hydrophilic stationary phase. The ion-pairs are then partitioned into the organic phase. Usually, porous silica gel is used as the support, and water or formamide is the stationary phase. The requirements are that a specific amount of counter-ions must be contained in the stationary phase, and the pH must be controlled to ionize the analytes. The eluent is not miscible with water, and counter-ions must stay in the stationary phase. The conditions were employed for the separation of sulfa drugs'' using tetrabutylammonium hydrogen sulfate (TBAHS04).The pH of this counter-ion is greater than 6 in water; it exists as tetrabutylammonium cation (TBAf) and sulfate anion (SO$-). n-Pentane was used as the major component of the eluent with a small amount of n-butanol, which enhanced the solubility of the pairedion. A small amount of water was added to the above mixture, which was then left standing until equilibrium was reached. A proportion of the paired-ion moved into the organic phase and a small amount of butanol and n-pentane

72

Chapter 4

moved into the aqueous phase. In this equilibrated system, the p H in the stationary phase is constant before and after equilibration. In the aqueous phase the amount of n-butanol is ca. 4 5 % (w/w) and that of n-pentane is negligible. The amount of water in the organic phase depends on the concentration of nbutanol. In 25% n-butanol-n-pentane (v/v), the water concentration in the organic phase is 1.5% (w/v), and 5.0% in the case of 50% (v/v) n-butanol-npentane. These values meet the above requirements; however, it is not simple in practice to obtain the appropriate experimental conditions.

Re versed-phase Ion-pa ir Liquid Chromat ograp hy

In this mode analytes form hydrophobic ion-pairs in the eluent, and the pairedions are retained on a hydrophobic stationary phase surface by their hydrophobicity. The hydrophobic paired ions are then eluted by reversed-phase liquid chromatography. Gradient methods, such as organic modifier gradients, counter-ion gradients, ionic-strength gradients, and pH gradients, can be applied in this system. Usually less than 0.1 wt% of the counter-ion and buffer concentration is employed. The selection of a stationary phase material is simple and a variety of hydrophobic stationary phase materials for reversed-phase liquid chromatography can be used. The organic modifier concentration depends on the hydrophobicity of stationary phase materials used and the ionpair reagent. Increasing the size of the counter-ion increases the retention time. The maximum retention time is reached when the counter-ion concentration reaches the micelle condition. The relationship between counter-ions and analytes is summarized in Table 4.4.

Example: ion-pair liquid chromatography of amino acids. Amino acids are zwitterions. The amino group can form an ion-pair with an alkanesulfonate ion (such as octanesulfonate), and the carboxyl group can form an ion-pair with a tetrabutylammonium ion, depending on the pH of the solution.
'

high pH

amino acids:

BYN+ -02C-CH2-NH2
systems

In reversed-phase liquid chromatography, the ionization of the solute decreases the retention. The addition of counter-ion under these conditions forces the formation of an ion-pair between the ionized solute and counter-ion, and then the retention of the analyte increases as the paired-ion is retained. Aspartic and glutamic acids were not retained in a buffer solution without a counter-ion; however, these acids were retained by the addition of octyl sulfate to the eluent, as seen in Figure 4.12. These amino acids can form a complex with copper ions and will be retained on the stationary phase. The addition of both a counter-ion and copper ions further increased the retention and made possible

Selection of the Eluent

73

Table 4.4 Relationship between analytes and counter-ions in ion-pair liquid chromatography
Analytes Strong and weak acids sulfate dyes carboxylic acids sulfa drugs Counter-ions Quaternary amines tetrabutyl ammonium tetramethyl ammonium hexadecyl trimethylammonium ammonia Tertiary amines trioctylamine Alkyl-, allyl-sulfonic acids pentane, hexane, heptane, octane, decane, dodecane sulfonic acids Phosphoric and sulfonic acids Alkyl sulfate, dodecyl sulfatea Perchloric acid

Strong and weak bases catecholamines

Basic compounds
a Use

as alkyl sulfonic acids, the selectivity is slightly different.

5.0
C .-

rn GIU Asp

oc
Figure 4.12

c u occcu Counter ion

Efect of counter-ions and copper on the retention of amino acids. Column, octadecyl-bonded silica gel, 25 cm x 4.6 mm i.d.; eluent, 0.01 M sodium acetate bufer (pH 5.6) containing 1.2 mM sodium octanesulfonate (Oc:)and/ M or 0.1 m copper acetate (Cu); flow rate, 1 ml min-'; detection, UV 220 nm. Compounds: Glu, glutamic acid; Asp, aspartic acid.

the separation of these compounds. Increasing the alkyl chain length of the counter-ion also enhanced the retention, as seen in Figure 4.13. Increasing the concentration of the counter-ion further increased the retention, but the retention reached a plateau, as seen in Figure 4.14. A total separation of the amino acids by reversed-phase ion-pair liquid chromatography could be performed." A column switching technique reduced the total analysis time (Figure 4.15).

74

Chapter 4

C .-

Ser

GlY

0-

C6

C7

C8

Figure4.13 Eflect of alkyl chain length o counter-ion on separation o amino acids. f f Column, octadecyl-bonded silica gel, 15 cm x 4.6 mm i.d.; eluent, 0.01 M sodium acetate bufler containing 0.4 M copper acetate and 1.0 mM sodium alkanesulfonate (pH 5.6); flow rate, 1 ml min-'; detection, UV 230 nm. Counter-ion: C6, sodium hexanesulfonate; C7, sodium heptanesulfonate; and C8, sodium octanesulfonate. Compounds: Ser, serine; Gly, glycine; Glu, glutamic acid; and Asp, aspartic acid.

1.6 3.2 6.4 mM 12.8 counter ion concentration

Figure 4.14 Efect o counter-ion concentration on separation of amino acids. Column, f octadecyl-bonded silica gel, 15 cm x 4.6 mm i.d.; eluent, 0.01 M sodium acetate bufler (pH 5.6) containing 0.4 mM copper acetate and 1.6-12.8 mM sodium heptanesulfonate;Jlow rate, detection, and analytes, see Figure 4.13.

Two mechanisms for retention in reversed-phase ion-pair liquid chromatography have been considered. One is the adsorption of the hydrophobic paired ion on the hydrophobic surface of stationary phase material. In the second mechanism, the hydrophobic counter-ion is held on the surface of the hydrophobic stationary phase, and the analyte ion is retained by ion-ion interactions, as shown in Figure 4.16. In the latter case, of a dynamic ion-exchange

Selection o the Eluent f

75

I .

4 .

10

20

30 min

Figure 4.15 Ion-pair liquid chromatography o free amino acids using a column switching f system. Column I , butyl-bonded silica gel, 50 x 4.6 mm i.d., 2, octyl-bonded silica gel, 50 x 4.6 mm i.d., and 3, octadecyl-bonded silica gel, 250 x 4.6 mm i.d.; eluent, 0.01 M sodium acetate bufer (pH 5.6) containing 4 m M copper acetate and 0.8 m~ sodium heptanesuljionate; Jow rate, 1 ml min-'; detection, UV235 nm. Peaks: 1, Tyr; 2, Val; 3, Met; 4,His; 5, Lys; 6, Ile, 7, Leu; 8, Phe; 9, Arg; 10, Asp; 11, Ser; 12, Glu; 13, Thr; 14, Gly; 15, Pro; 16, Cys; and 17, Ala. 1-9 were separated on column I and 10-1 7 were f separated by a combination o columns 2 and 3.

Figure 4.16 Model of reversed-phase ion-pair liquid chromatography in which counterions are held on the stationary-phase surface.

mechanism, the column efficiency is much higher than in the corresponding ionexchange liquid chromatography using an ion-exchange resin as the stationary phase. Examples of samples, counter-ions, and stationary phases are summarized in Table 4.5.

p H Effects
The retention of acids in reversed-phase liquid chromatography can be predicted by Equation 4.1. The same equation can be applied to predict the

Table 4.5 Applications of reversed-phase ion pair liquid chromatography

Stationary phase material Eluenl Ref-

Sample

Counter-ion

ODS silica ODS silica

ODS silica NH2 or OH silica Phenyl silica C8 silica Me silica C18 silica Polystyrene Silicone/cellulose C8 silica

50 mM TMA in 20% MeOH, pH 6.1 60% aq. MeOH

24 17 18
45

Alkaloid TMA Alkanesulfonic acid Cetylpyridium Tris(2,2-dipyridyl)ruthenium(11)mono dicarboxylic acids & Alkylsulfonic acid Amine Tetrabromophenolphthalein Amino acid Naph t halene-2-sulfododecanesulfonic acid Dodecyl sulfate Aminoglycoside TFA PFPa + TFA Aminop henol Bis(2-ethylhexy1)phosphate Antidepressive amine Trimet hylnonylammonium Ph silica

14 15 62 38 39 12 57

Apomorphine

Dodecyl sulfate

35
473 56 19

Anions Ascorbic acid Barbital

TBA Tridecylammonium TBA

0.01 M Borate buffer pH 10.0 Dichloromet hane/MeOH 50 mM Phosphoric acid 20% ACN + 50 mM H2S04,5 mM NaHS04 Na citrate + Na sulfate + PrOH + SSD 0.2 M TFA PFPa + TFA Citric buffer pH 3.8 NaBr + trime thylnony 1 ammonium/Pentanol, phosphate buffer 20 mM NalHP04 + 30 mM citric buffer pH 3.25 + MeOH + ACN (55:36:9) TBA + salicylic acid/pH 4.62 Tridecylammonium, pH 5.0 0.03 M TBA pH 7.7

Benzoates Benzoic acid derivatives Carboxylic acid

Catecholamine

TBA TBA 1-Phenetyl-2-pyrroliniumcetride Octyl sulfate

C18 silica C 18 silica Bu tylnitrile/C2 silica Silanized silica C2 silica Phenyl silica C8 silica C18 silica

65 19 14 15 34

Q
41,54 g

Cefaclor/cefradine

TBA

C 18 silica

PeOH -t- phosphate buffer 0.03 M TBA pH 7.4 Acetate buffer pH 4.6 35% ACN + 0.05 M ammonium acetate 50 mM Citric acid + 50 p~ EDTA in 50 mM Na2HP04 (2 :1) Phosphate buffer pH 7.0 + TBA + ACN

5
.a,

3
+ s

s'
Q

C18 silica C 18 silica C8 silica C18 silica C18 silica

5 mM Dodecyl sulfate, 0.1 % acetic acid in 50% ACN 70% MeOH + 50 mM NH4C02H NaC104/ACN Dodecyl sulfate/phosphoric acid + ACN Heptanesulfonate/phosphate buffer ACN
23 16 48 50 60 66 14 20 59 16 37 33 19 49 21 36 46 42 32 27 63 16 43 21 HSA + MeOH + acetic acid 50 m Phosphoric acid M 5 mM TBA/45% MeOH Octanesulfonate + citric acid/phosphate buffer, MeOH 70% MeOH + NH40H pH 4.8 Octanesulfonate + ninhydrin + ACN + MeOH 5 r n TBA + ACN (15:85) ~ 0.03 M TBA pH 7.4 TBA/MeOH + phosphate buffer Pentanesulfonate + NaH2P04/ACN pH 3 2 mM TBAP + ACN/Phosphate buffer pH 7.0/gradient Fe(phen)3S04/0.5 mM acetate buffer 3 mM Phosphate buffer pH 4 + octanesulfonate/ACN 1 g Heptanesulfonate + 150 mg EDTA/1 1 water pH 4.0 50 mM NH4COZH/SO mM TFA, pH 1.3 Phosphate buffer + PeOH 30% MeOH + 20 mM N H 4 0 H pH 6.4 TBAP/ACN + perchloric acid TBA + MeOH + water + acetic acid (0.15 :62 :35 :3)/gradient

Choline/3,5-dini tro-benzoic acid derivatives Dodecyl sulfate Cocaine Ammonium Cosmetic dyes NaHCI4 Daunorubicin Dodecyl sulfate Diazonium salts Heptanesulfona te L-3,4-Dihydroxyp hen y lalanine metabolites Heptanesulfonate Dipeptide Naphthalene-2-sulfonate Dyes TBA Enkephalins Octanesulfonate Erythromycin Ammonium Guanidino compounds Octanesulfonate Heroin TBA Indoleacetic acid TBA C 18 silica Ph silica C18 silica C 18 silica C 18 silica C8 silica PrNH silica Pentanol/C2 silica C 18 silica C18 silica C 18 silica Polystyrene C8 silica C 18 silica C18 silica Silanized silica C 18 silica C8 silica C l 8 silica

trl 2

'

Iodine, Thiosulfate TBA Neostigmine, pyridostigmin, edrohonium Pentane sulfate Oligodeoxyribonucleotides TBA Organic acids Iron(rr) 1,lO-phenanthroline Oxymorphone Octanesulfonate Penicillamine Heptanesulfona te Peptide Ammonia, TFA Phenylacetic acid TBA Phen ylenediamine Ammonium Pirlimycin TBAP Prostaglandin E l TBA

(continued)
4 -4

Table 4.5 Continued

Stationary phase material
C 18 silica C18 silica Phenyl silica C18 silica
C 18 silica

Sump Ie

Counter-ion

Eluent

Ref.

Pyrimidine, Purine, Azapurine metabolite TBA Pyrocarbine, Phyzostygmin Pentanesulfonate Quaternary amine 1-Phenetyl-2-pyrrolinium Rebamisol Heptanesulfona te C8 silica Butylni trile/C2 silica SAS silica Silanized silica SC-TAS C 18 silica C 18 silica

22 29 14 30 28
31

Riboxamide Salbutamol

HTAB Heptanesulfonate

Sulfonamide

TBA

10 mM TBA, pH 8.0 in 0.2% ACN 5 mM Heptanesulfonate, pH 3.6 in 4% MeOH Acetate buffer pH 4.6 2% Heptanesulfonate in 0.2% acetic acid + MeOH (55 + 4) 1 mM HTAB in 10 m phosphate buffer pH 6.0 M 5 mM Heptanesulfonate in 5.8 m phosphate buffer M pH 6.0 + MeOH (3 1: 69) 0.03 M TBA pH 7.9 PrOH/H20 or PrOH + CH2C12/H20 PeOH + phosphate buffer ACN TBA/formic acid + NaOH Octanesulfonate + N,N-dimethyloctylamine/MeOH

19 64 65 61 59 51 26 40
53 44 18

Sulfonic acids Sulfonates Tetracycline Tartrazine Tetrahydrozoline Theophylline metabolites

Cetyltrimethylammonium TBA HC104 TBA Octanesulfonate

Tricyclic antidepressant Tris(2,2-bipyridyl)Ru N,N-Trime thy lene-bis

TBA TBA Decylamine Alkylammonium Heptanesulfonate Heptanesulfonate

C 18 silica C18 silica C 18 silica C8, C18 silica C 18 silica C18 silica

TBA, pH 4.75 in 50% MeOH, gradient TBA + 15 mM Tris buffer + ACN + MeOH Decylamine/ACN MeOH/phosphate buffer pH 2 5 mM Heptanesulfonate buffer pH 3.5 Heptanesulfona te/ACN

58

(p y ridium-4-aldoxime)

dibromide

Selection of the Eluent

79

retention factor in reversed-phase ion-pair liquid chromatography. The pH and counter-ion effects on the chromatographic behaviour of indoleacetic acid are demonstrated in Figure 4.17. The addition of a surfactant counter-ion reduced the retention factor at low pH due to the surface modification of the stationary phase material. Covering the surface of the stationary phase with the surfactant reduces the hydrophobicity of the stationary phase material. The addition of the tetrabutylammonium counter-ion increased the retention factor at high pH. The pK, of the indole acetate was 5.15 without surfactant, 4.85 with octyl sulfate ion, and 5.60 with tetrabutylammonium ion. That is, the addition of a same-charged hydrophobic ion reduced the PKa value, and the addition of the counter-ion increased the pK, value. The difference in the PKa value on the addition of surfactant is not constant; it is affected by the kind of ion and the concentration. It is difficult to estimate the PKa change. The retention factor of the molecular form of the analyte, k,, and the retention factor of the paired-ion7kip, can be easily measured experimentally. When the retention factor is measured at low and high pH, and the pK, value is known, k at a given pH is roughly estimated from the following equation:

where Ka is the dissociation constant of the ion-pair complex.

15

NaP

10

0
2 5

10

PH
Figure 4.17 Selectivity of counter-ion for the retention of indoleacetic acid. Column, Hitachi 3011 (polystyrene gel), 25 cm x 4.1 mm i d . ; eluent, 50 mM sodium phosphate ( N a p ) containing 25 mM tetrabutylammonium (TBA) or octyl sulfate (C8) ion in 20% acetonitrile solution; flow rate, 1 ml min-'; temperature, 40 "C; detection, UV 254 nm.

80

Chapter 4

Effect of the Ionic Strength of the Buffer Solution

In ion-exchange liquid chromatography, an increase in the buffer concentration or the addition of a salt is effective in improving the resolution. Increasing the inorganic ion concentration reduces the ion-pair formation in normal-phase ion-pair partition liquid chromatography. The retention factor values then become smaller. In contrast, the buffer concentration does not affect the retention of the paired-ion in reversed-phase ion-pair liquid chromatography. However, the solubility of inorganic salts in organic solvent-rich eluents is poor, making the maintenance of the instrument difficult. It is highly recommended that a low-concentration buffer solution should be used if the pH is not changed during the separation.

Selectivity of Organic Modijier and the Concentration Effect

The selectivity and the concentration effect of the organic modifier in the eluent are the same as those explained above for reversed-phase liquid chromatography. Increasing the modifier concentration reduces the retention factor, as does decreasing the counter-ion concentration. The degree of change in the retention factor is due to changes in the solubility of the solute in the organic solvent.

Column Temperature Effects

Increasing the column temperature reduces the retention factor. The ion-pair formation is based on a chemical equilibrium; therefore, temperature control is important to obtain reproducible results.

Detection
The molecular absorption intensity of polar compounds is usually small, but highly sensitive detection can be obtained after pre- or post-column derivatizations. The use of ultraviolet absorption or fluorescence-active counter-ions makes it possible to achieve highly sensitive detection of polar compounds and enhance the capability of ion-pair liquid chromatography. For example, N,Ndimethylprotriptyline has been used as a counter-ion for carboxylic acids12 and picric acid for quaternary amines13 in normal-phase ion-pair partition liquid chromatography. Phenethylammonium, cetylpyridinium, 1-phenethyl-2-pycolinium, and naphthalene-2-sulfonic acid have been used for sulfonic acid and alkyl amines detection. l 4 ? l 5 Ion-pair post-column extraction was applied on-line for fluorescence detection. l 6 Ion-pair liquid chromatography can be applied to compounds separated by ion-exchange liquid chromatography, and mixtures of ionic and non-ionic compounds are easily separated. The latter separation is difficult by ionexchange liquid chromatography. Anions can be separated by reversed-phase ion-pair liquid chromatography (Figure 4.18).

Selection of the Eluent

81

I' 0

min

I 30

Figure 4.18 Analysis o anions in water using ion-pair liquid chromatography. Column, f octadecyl-bonded silica gel, 15 cm x 4.6 mm i.d.; eluent, 2 mM tetrabutylammonium hydroxide (pH 5.3) in 3% acetonitrile-water; flow rate, 1 ml min-'; detection, UV 200 nm. Peaks: 1, Br-; 2, NO3-; and 3, I - .

Trouble-shooting in Ion-pair Liquid Chromatography

Ion-pair liquid chromatography can be applied to the separation of a wide variety of compounds. However, the flexibility of the selection of eluent components can confuse the operation. Trouble-shooting solutions are summarized in Table 4.6.

3 Normal-phase Liquid Chromatography

Normal-phase liquid chromatography was formerly called adsorption liquid chromatography. Pure or mixed organic solvents are used as the eluent and stationary phase adsorbent is more polar than the eluent. Hydrogen bonding is one of the important molecular interactions between sample molecules and the adsorbent. The possible molecular interactions are summarized in Table 1.1. When no molecular interaction is recognized, such chromatography is called size-exclusion liquid chromatography. The basic molecular interaction in normal-phase liquid chromatography is electrostatic forces. The sample molecules are retained strongly by hydrogen bonding when the sample molecules themselves or the adsorbent act as both a hydrogen-bond acceptor and donor. The specificity of normal-phase liquid chromatography is the result of the direct formation of a strong molecular interaction between sample molecules and the adsorbent. The position of substitution of sample molecules directly affects the separation of isomers. The substituent effect is weak in reversed-phase liquid chromatography. Therefore, normal-phase liquid chromatography is suitable for the separation of isomers, such as cisltrans, ortholmetalpara, and steric isomers. The steric effect is especially important for chiral separations on suitable chiral columns. A

Table 4.6 Trouble-shooting for ion-pair liquid chromatography

Solution

Trouble

Reason

No peak found

Low conc. sample soln. Unsuitable detector

Disappeared in ghost peak (solvent, coun ter-ion) Adsorption in column

No retention of analyte

Leak from injector Poor flow rate due to surfactant Poor detection due to surfactant Unsuitable column Unsuitable counter-ion

Non-reproducible retention time

Unsuitable conc. organic modifier Too high conc. buffer Unsuitable pH Unstable flow rate Unstable temperature Non-equilibrated column Too diluted conc. buffer

Check sensitivity of detector & adjust conc. sample soln. Usually UV absorption of eluent is high because of counter-ion and buffer Select suitable wavelength or change detector Increase conc. sample soln. & compare with previous chromatogram. Or measure diode array chromatogram Change column Increase polarity of column Increase conc. organic modifier Dilute conc. counter-ion Clean and fix injector Clean and fix pump head Clean and fix detector flow cell Change column Select further hydrophobic counter-ion Increase conc. counter-ion Decrease conc. organic modifier or dont use organic modifier Decrease conc. buffer, especially inorganic buffer Change pH, take care with pK, of analytes Clean pump and flow line Use column heater or oven Equilibrate column, especially for initial injections Change buffer

Peak tailing

Peak broadening

Noise-like peak

Change column Increase conc. counter-ion Change pH suitable for ionization of analytes Equilibrate column, usually for 30-120 min Use moderate gradient Reduce amount injected Increase conc. organic modifier Change column Use more hydrophobic counter-ion Change pH of eluent Change column, use guard column Change column to suitable one, change eluent Clean pump head Clean injector

High pressure drop

Too strong adsorption on column Too dilute counter-ion Unsuitable eluent pH Non-equilibrated column Due to gradient Overloaded sample Low conc. organic Damaged column Unsuitable counter-ion Unsuitable pH High conc. basic counter-ion pH > 7.5 Unstable flow rate (air, recrystallized salt, and counter-ion in pump head) Recrystallization of salt counter-ion

Low sensitivity

Due to injector or pump Too high conc. buffer Unsuitable wavelength Dirty flow cell Unsuitable counter-ion

Aged lamp or detector

Change eluent, buffer components Wash column with water then methanol Wash and fix injector and pump Dilute conc. buffer and counter-ion Select appropriate wavelength Wash flow cell with water and methanol Change counter-ion to one having a chromophore or fluorophore and change the condition Fix detector

w 03

84

Chapter 4

variety of chiral separation phases have been synthesized; however, one phase cannot separate many different types of optical isomers, because the steric effect also depends on the molecular structure of the analytes. In ligand-exchange liquid chromatography, the separation is performed by the replacement of analytes, which form a complex with stationary phase materials or a reagent trapped on the stationary phase surface, by components in the eluent, which act as similar ligands to the analytes. In charge-transfer complex liquid chromatography, analytes are retained by charge-transfer complex formation on the surface of stationary phase material and are replaced by other components in the eluent which form a charge-transfer complex with the stationary phase materials or with reagents trapped on the stationary phase surface. Normal-phase liquid chromatography is thus a steric-selective separation method. The molecular properties of steric isomers are not easily obtained and the molecular properties of optical isomers estimated by computational chemical calculation are the same. Therefore, the development of prediction methods for retention times in normal-phase liquid chromatography is difficult compared with reversed-phase liquid chromatography, where the hydrophobicity of the molecule is the predominant determinant of retention differences. When the molecular structure is known, the separation conditions in normal-phase LC can be estimated from Table 1 . 1 , and from the solvent selectivity. A small-scale thin-layer liquid chromatographic separation is often a good tool to find a suitable eluent. When a silica gel column is used, the formation of a monolayer of water on the surface of the silica gel is an important technique. A watersaturated very non-polar solvent should be used as the base solvent, such as water-saturated n-hexane or isooctane.
Example I : Chromatography o phthalic acid alkyl esters. Phthalic acid alkyl f esters have been chromatographed on different types of stationary phase materials in normal-phase liquid chromatography. The stationary phase materials used were dimethylsilylated silica gel, propyl cyano-bonded silica gel, polystyrene gel, pellicular-type silica gel, and pellicular-type anion exchangers. In each case the longer the alkyl chain length of the phthalate, the shorter was the retention time. This result suggested that the elution order depended on the alkyl chain length of the samples and was not affected by the stationary phase materials, and that molecular interaction and solubility, not the type of stationary phase material, were important for the ~ e p a r a t i o n . ~ ~ Example 2: Chromatography o nitroaniline isomers. The elution order of the f nitroaniline isomers was ortho, meta, and para in normal-phase liquid chromatography using n-butanol-n-hexane mixtures as the eluent, when the stationary phase material was either silica gel, alumina, an ion-exchanger, polystyrene gel, or octadecyl-bonded silica gel. The results indicate that the separation of these compounds can be performed on a range of different types of stationary phase materials if the correct eluent is selected. The best separation will be achieved by the right combination of stationary phase material and eluent.68

Selection of the Eluent

85

Classification of Solvents
The most practical method for finding a good eluent for normal-phase liquid chromatography is a trial experiment using thin-layer chromatography (TLC). When a rough separation is achieved by TLC and the maximum Rfvalue is less than 0.5, the developing solvent is a good candidate to be a suitable eluent for column liquid chromatography. The modification of components depends on the surface activity, i.e. the water content, and the surface area of silica gel. The selection of the solvent can be based on the solvent properties classified by polarity (Table 4.7), hydrogen bonding (Table 4.8), and solubility parameters (Table 4.9). Snyder has classified solvents according to their hydrogen bonding acceptor, hydrogen bonding donor, and dipole properties. 70 The solvents useful for liquid chromatography are shown in Table 4.9. Water and chloroform are classified in the same group (VIII); however, these solvents are not miscible. The method for identifying a good developing solvent for TLC is also applicable to liquid chromatography. First a solvent in which the analyte is very soluble, based on the concept of 'like dissolves like' developed by Freizer, must be selected, then a solvent in which the analyte is not soluble is selected and these solvents have to be miscible together. The eluent is thus a mixture of solvents for the analyte. The selection of a solvent is based on the solvent strength' E experimentally obtained on silica gel and alumina.69 An eluent having E' = 0.40 on alumina can be made from a mixture of methyl acetate-pentane (19 : 8 1) or acetone-pentane (28 : 72). An eluent having EO = 0.20 on alumina can be made from a mixture of acetone-pentane (6 :94), chloroform-pentane (1 5 : 85), or benzene-pentane (28:72), as shown in Figure 4.19. However, no linear relationship between

Table 4.7 ClassiJication of solvents by Hecker7'

Non-polar solvent Hexane Heptane Isooctane Cyclohexane Cyclopentane Carbon disulfide Carbon tetrachloride Benzene Trichloroethylene Dichloroet hy lene Chloroform Tetrachloroethylene Dichlorome t hane Diethyl ether Methyl ethyl ketone Ethyl acetate Pyridine Tetrahydro furan Dioxan Acetone Aniline Phenol Ethanol Glycol monomethyl ether Acetic acid Acetoni trile Nitromethane Formic acid Morpholine Formamide Water Inorganic acid Salt, Buffer Polar solvent

86

Chapter 4

Table 4.8 Class$cation o solvent based on hydrogen bonding f

Strong donor & acceptor Weak donor & acceptor Donor Acceptor Water, glycol, glycerol, aminoalcohol, hydroxylamine, polyphenol, and amide Alcohol, organic acid, phenol, primary and secondary amines, oxime, nitro and nitrile compounds having hydrogen at a-position, hydrazine, HF, and HCN Ether, ketone, aldehyde, ester, tert-amines, pyridine, and nitro and nitrile compounds not having hydrogen at aposition Chloroform, dichloroethane, CH3CHC12, CHzClCH2C1, and CH2CICHClCH2CI Hydrocarbons, carbon disulfide, mercaptans, solvents not included in above groups

No hydrogen bonding

5 m

z U J
w
I

r a

A 0 A0 A0 0

0
0

0.2-* 3 go0
#D

c Q) >
0

09

' I

Figure 4.19 Solvent strength of combination of n-pentane and more polar solvents in normal-phase liquid chromatography using alumina. Symbols: A,methyl chloroform; and 0, benzene. acetate; 0 ,acetone; 0,

these mixed solvents is obtained; this is both a rough estimation method and a practical method. The order of solvent strength (6) (Table 4.9) and the adsorption parameter on alumina ( are not parallel. This indicates that solvent strength cannot be E ' ) directly related to solvent selection in liquid chromatography. One way of using Table 4.9 in practice is to rearrange the solvents in terms of the values of E'. Increasing the E values strengthens the solubility. The retention time thus ' becomes shorter. This method is not perfect, and the elution power also depends on the structure of the analytes. Another method is the use of 8d,a a and a h . , , These values indicate the interaction properties of the solvents. The solvent can be selected on the basis of these values and the properties of the analytes. The sum of these values and the steric effect may affect the chromatogram.

Table 4.9 Solvents commonly used in liquid chromatography

6,
&O

%
2
ul

2
6h

Solvent

dd

60

vwv
M,
dm

RI
UV

BP

("C) q
5 rp
-

5.

t?
i $

0 0 0 0 0 0 0 0 0 0
-0.25 0.01 0.28 0.00 0.0 1 0.0 1 0.38 0.54 0.05 0.04 0.30 0.18 0.58
-

1 0 0.5 0 0 0 2 3.5 0 0 0 0.5 2 6.5 0.5 2 0.5 2.5 0.5 4 5 5

0 0 0 0.5 0 0 0 0 0 4 0 5
-

1.25 1.3914 1.368 1.3575 1.3749 1.3876 1.3524 1.399 1.4064 1.4262 1.460 1 1.460 1 1.3724 1.385 1.4458 1.362

0 0 0.5 0.00 0 0.0 1.15 0 0.00 0 3 0 1.88 4 1.15 4.5 0 2.69 1.14 3.09 2.37 4.0 3.86
-

For normal-phase liquid chromatography Polyfluoroalkanes 6.0 6.0 Isooctane 7.0 7.0 Diisopropyl ether 6.9 7.0 n-Pentane 7.1 7.1 n-Hexane 7.3 7.3 n-Heptane 7.4 7.3 Diethyl ether 7.4 6.7 Triethylamine 7.5 7.5 Cyclopentane 8.1 8.1 Cyclohexane 8.2 8.2 Propylchloride 7.3 8.3 Carbon tetrachloride 8.6 8.6 Ethyl acetate 7.0 8.6 Propylamine 8.7 7.3 Chloroform 8.1 9.1 Methyl acetate 9.2 6.8 Benzene 9.2 9.2 6.8 Acetone 9.4 Dichloromethane 6.4 9.6 Propanol 7.2 10.2 Pyridine 10.4 9.0 Ethanol 6.8 11.2 Dimethylformamide 7.9 11.5 Acetic acid 12.4 7.0
-

0.40 0.60 0.32 0.56 0.42 0.82 0.71 0.88 1.o

0.50 0.37 0.23 0.313 0.41 0.24 0.36 0.44 1.00 0.35 0.97 0.45 0.35 0.57 0.37 0.65 0.36 0.44 2.3 0.95 1.20 0.92 1.26 1.5011 1.3587 1.424 1 1.3856 1.5102 1.361 1.4305 1.372

101.93 82.51 67.02 77.30 89.70 58.98 86.57 59.39 70.89 51.18 53.86 57.15 50.14 44.50 45.29 54.47 40.60 36.14 46.13 50.39 34.24 49.42 32.87

114.230 102.176 72.150 86.177 100.203 74.122 101.191 70.134 84.161 78.541 153.823 88.106 59.111 119.378 74.079 78.113 58.080 84.933 60.096 79.101 46.069 73.094 60.052

0.01 5 1.266 0.020 0.016 0.006 1.557 0.848 0.003 0.000 1.725 0.001 1.769 1.444 1.156 1.725 1.305 2.91 1 1.503 1.534 1.973 1.551 3.550 1.885

210 210 220 190 195 200 218 345 200 200 263 263 256 245 260 278 330 233 210 330 210 268 230

-50 98 68 33 69 98 34 90 49 80 76 76 76 48 60 56 80 56 40 98 115 78.5 153 118

(continued)
00

Table 4.9 Continued

00 00

Solvent

0 4 1.75 2.69 0.45 3.09

-

0 0.5 0 0 0 4
5 -

4.0

3.44 3.9 2.87

3.5 6.5 3 2.5 3 4 5 2.5 5 7.5

.o
2.24 0.55 -

0.36 0.35 0.55 0.36 1.37 2.3 1.9 1.20 1.26 0.38 1.26

0 7.5
-

0.54 0.45 0.56 0.56 0.82 0.82 0.88 1.o 0.65 1 0.6 0.95 -

For re versed-phuse liquid chromatography Triethylamine 7.5 7.5 7.3 Propylamine 8.7 9.1 7.5 Tetrahydrofuran 9.4 6.8 Acetone 9.8 7.8 Dioxane 10.2 7.2 Pro panol Isopropanol Ethanol 11.2 6.8 7.9 11.5 Dimethyl formamide Acetoni trile 11.8 6.5 12.4 7.0 Acetic acid Trifluoroacetic acid 12.8 8.4 Dimethyl sulfoxide 12.9 6.2 Methanol 8.3 13.5 Ethanolamine Formamide 8.3 17.9 6.3 21.0 Water

L L L
L L L
1.oo

L L L

86.57 50.14 40.87 40.60 56.16 46.13 45.66 34.24 49.42 27.68 32.87 37.75 48.43 22.40 41a.57 24.42 9.38

101.191 59.1 11 68.075 58.080 88.106 60.096 60.096 46.069 73.094 41.052 60.052 114.024 78.129 32.042 6 1.083 45.041 18.015

0.848 1.444 0.492 2.91 1 0,010 1.534 1.618 1.551 3.550 2.894 1.885 1.879 3.948 1.621 1.166 3.699 1.860

1.399 1.385 1.4072 1.3587 1.4224 1.3856 1.384 1.361 1.372 1.3441 1.372 1.480 1.483 1.3284 1.4540 1.4470 1.333

345 212 330 215 210 205 210 230 190 230 260 260 205

275 190

90 48 190 56 101 98 82 78.5 118 82 118 72.4 190 65 170 210 100

~~

6: Total solubility parameter indicates solvent strength and polarity. Larger 6 means polar, smaller means non-polar (hydrophobic), Sd: the ability of a solvent to participate in dispersive interactions, which indicates degree of solubility of aromatic compounds having halogen and sulfur substituents. Larger values mean strong dispersive interaction. 6,: orientation interaction; indicates degree of solubility of dipole compounds. 6,: proton donor property; indicates degree of solubility of alcohol, phenol, and carboxylic acid. &: proton acceptor property, indicates selective degree of sqlubility of basic compounds. 8': solvent strength in normal-phase liquid chromatography using alumina. q: viscosity (cp at 20 "C), VWV*: van der Waals volume A3, M,: relative molecular mass, d,: dipole moment in debye, RI: refractive index, UV: UV cut-off (nm), BP: boiling point ("C), L: too large, *calculated using CACheTMprogram from Sony-Tektronix. This table is based on references by L.R. Snyder (ed.), 'Principles of adsorption chromatography', Marcel Dekker, New York, 1968, p. 174, and J.J. Kirkland (ed.), 'Modern practice of liquid chromatography', Wiley-Interscience,New York, 1971, p. 136.

2 s 9
A

Selection of the Eluent

89

For example, if the separation of a steroid mixture is difficult in one chromatographic system, the selection of a different system is needed to separate the mixtures. The substituents with a hydrogen-bonding effect due to a hydroxy group, a hydrogen-bonding donor effect of a carbonyl group and a hydrophobic effect due to an ester suggest the elution order and solvent selection. That is, the solvent selection is based on &, So, a and Sh of analytes. Subsequently, another , solvent that is miscible with the first solvent is used for dilution. The final eluent is a mixture of these two s ~ l v e n t s . ~ l - ~ ~

Preparation of the Eluent

Like dissolves like is the basic concept for the selection of solvents in the eluent for liquid chromatography. Controlling the solubility of analytes is the key to success. If the selected solvent or mixture of solvents does not interfere with detection, it is a good eluent. The selection of a suitable solvent for lowwavelength absorption detection and post-column derivatization detection is important to obtain highly sensitive detection. The selection of a volatile solvent is the key for preparative-scale liquid chromatography and for mass spectrometric detection. The selection of the solvent is based on the retention mechanism. The retention of analytes on stationary phase material is based on the physicochemical interactions. The molecular interactions in thin-layer chromatography have been extensively discussed, and are related to the solubility of solutes in the solvent. The solubility is explained as the sum of the London dispersion (van der Waals force for non-polar molecules), repulsion, Coulombic forces (compounds form a complex by ion-ion interaction, e.g. ionic crystals dissolve in solvents with a strong conductivity), dipole-dipole interactions, inductive effects, charge-transfer interactions, covalent bonding, hydrogen bonding, and iondipole interactions. The steric effect should be included in the above interactions in liquid chromatographic separation. One example is the separation of tricarboxylic acid cycle organic acids. These organic acids were originally separated on sulfonic acid-treated silica gel as the stationary phase material, and a chloroform and isopentyl alcohol mixture was used as the eluent. However, this eluent is not suitable for UV detection. A mixture of n-hexane, THF, and tert-butanol was therefore selected for the same ~eparation.~~ One component of the eluent should have properties similar to those of the analytes, and this solvent is diluted by another solvent to control the retention time. The basic idea can be understood from the chromatographic behaviour of phthalic acid esters and polycyclic aromatic hydrocarbons (PAH). This approach can be applied to the separation procedure for a variety of stationary phase materials, including silica gel, polystyrene gel, and ionexchangers. The elution order of phthalic esters is related to the carbon chain length. The longer the chain length, the shorter the retention time in normal-phase liquid chromatography, and the elution order is reversed in reversed-phase liquid

90

Chapter 4

chromatography. The elution order of PAHs is based on the number of double bonds, an increased number of double bonds leads to longer retention time in both normal- and reversed-phase liquid chromatographies. These elution orders are always observed on a variety of stationary phase materials, including ionexchangers. The retention time can be modified by changing the eluent components or the ratio of the solvent mixture. Increasing the solubility in the eluent shortens the retention time in both normal and reversed-phase liquid chromatographies. Increasing the alkyl chain length of analytes makes them more similar to alkanes. The analyte becomes more hydrophobic (lipophilic) and hence soluble in n-hexane, but not in water. The solubility of both short and long alkyl chain compounds is good in ethanol, which is miscible with n-hexane and water. The addition of ethanol to the eluent makes the retention time shorter in both normal and reversed-phase liquid chromatographies. This phenomenon can be observed in the chromatographic behaviour of aliphatic acids. However, the solubility of PAHs is poor even in ethanol; therefore, the elution order is always the same in both normal- and reversed-phase liquid chroma tographies. In normal-phase liquid chromatography, the elution order of benzene derivatives on silica gel is alkylbenzenes with a long alkyl chain < alkylbenzenes with a shorter alkyl chain < benzene < fluorobenzene < chlorobenzene < cyanobenzene < nitrobenzene < aniline < phenol < benzoic acid. The elution < order of aniline derivatives is dinitroaniline < benzene < anisidine < chloroaniline < nitroaniline < phenylenediamine < aminophenol. This elution order is also observed for anion-exchangers. Alkanes are almost unretained. Alkenes are retained according to the number of double bonds. The elution order for cis and trans isomers is cis,cis < cis,trans < trans,trans. This also suggests that the eluent component should be selected according to the 6, of dipole moment, 6, of hydrogen bonding acceptor, and 6 h of hydrogen bonding donor, but not the polarity 6. In other interactions the surface silanol groups of silica gel can form hydrogen bonds, and an alumina surface can form hydrogen bonds and a charge-transfer complex. However, such molecular interactions are caused by positive and negative charge sites, as in Lewis acid-base interactions. The chloro-ion of chloride-form anion-exchangers may form hydrogen bonds with aniline and phenol. Silver ions of silver-form cation-exchangers form charge-transfer complexes with the n-electrons of alkenes and aromatic compounds.

Size-exclusion Liquid Chromatography (SEC)

This separation method is based on the molecular size of analytes. Analytes pass through porous stationary phase materials having different pore sizes, and molecular interactions between analytes and the stationary phase surface must be eliminated. A very strong solvent is therefore required in this system. This system is also called gel filtration liquid chromatography, gel-permeation liquid chromatography, or molecular sieve chromatography. This system is used to

Selection of the Eluent

91

measure the average molecular mass and for the purification of large- from small-size molecules.

Aqueous Phase Size-exclusion

The elution time is related to the logarithmic relative molecular mass (log M,) within a certain range of M , for a particular column, as seen in Figure 4.20. The relationship for proteins is not as simple as that for dextran and polyethylene glycol. This is because the molecular shape of proteins differ even though 0.2 M sodium phosphate buffer, pH 6.8, was used as the solvent. The molecular shape effect can be understood from the different calibration curves in Figure 4.20. The elution time of polyethylene glycol, whose structure is like a straight chain, was longer than that of dextran. Dextran and polyethylene glycol can be eluted with only water as the solvent. The hydrolysates of dextran have been separated by SEC in pure water.75 The non-uniform behaviour of proteins makes the measurement of their molecular mass difficult; however, their rough estimation and purification are very important for biochemical research. Generally, phosphate buffer and trisHCl buffer (pH 7) are used for SEC for biological polymers, and some additives are required for further separation. Only 0.2 M sodium phosphate buffer was

. I

E 0
&
0
Q)

105

lo4 103

3
I

20 30 Elution time
Figure 4.20

40 min

Calibration curves for size-exclusion liquid chromatography. Column, TSK GEL G3000SW, 120 cm x 7.5 mm i.d.; eluent, 0.2 M sodium phosphate bufler pH 6.8; flow rate, 1.0 ml min-'. Standards: 1, protein; 2, dextran; and 3 , polyethylene glycol. (Reproduced by permission from Tosoh data)

92

Chapter 4

used for the analysis of gelatin. However, 0.1 M sodium phosphate buffer (pH 7.0) containing 0.1 M sodium chloride was required for the analysis of ribonucleic acid from E. ~ o l iand 0.1 M sodium phosphate buffer (pH 7.0) containing ,~ 0.3 M sodium chloride was required for that of serum from bone marrow.* TrisHCl buffer (pH 7.5) containing 0.1% sodium dodecylsulfate was used for the analysis of cellular membrane protein from rat micro some^,^ and 50 mM TrisHCl buffer (pH 7.2) containing 2 mM EDTA and 300 mM NaCl for the analysis of plasmid DNA fragments as the elution solvent.8 SEC is a useful tool for monitoring enzyme reactions, as seen in Figure 4.21, where the speed of decomposition of P-lactoglobulin by a-chymotrypsin is shown. SEC is widely used for purification of proteins, but the separation is due only to the difference in molecular mass. Therefore, ion-exchange liquid chromatography is combined with SEC to improve the selectivity.

Organic Phase Size-exclusion

The average molecular mass of synthesized polymers can be measured by SEC. This system has been called gel-permeation liquid chromatography, using porous polystyrene gel as the stationary phase material. Polymers are usually dissolved in THF or chloroform for separation by SEC. The molecular density

I 10

I 20

min

30

f Figure 4.21 Monitoring o an enzyme reaction using size-exclusion liquid chromatography. Column, TSK GEL G3000SW, 60 cm x 7.5 mm i.d.; eluent, 0.07 M potassium phosphate bufler containing 0.1 M potassium chloride; flow rate, 1 rnl min-'; detection, UV 280 nm. Peaks: 1, J-lactoglobulin; 2, a-chymotrypsin; and 3, decomposedproducts. (Reproduced by permission from Tosoh data)

Selection of the Eluent

93

10

15

min 20

25

Figure 4.22 High temperature size-exclusion liquid chromatography of an engineering plastic, poly(pheny1 sulfate). Column, SSC GPS-3506, 50 cm x 8 mm i.d.; eluent, I-chloronaphthalene; $ow rate, 1 .O ml min- column temperature, 2 10 "C; detector, refractive index detector. (Reproduced by permission from Senshu Kagaku data)

';

of analytes is measured by comparison of chromatograms obtained from light scattering and refractive index detectors. Newly synthesized organic polymers having high temperature resistance and physical strength are not soluble in THF or chloroform. High-temperature SEC is therefore being developed, and the pressure-drop measurement detector and flame ionization detector are also used for monitoring the high temperature effluent. A chromatogram of an engineering plastic, poly(pheny1 sulfate), is shown in Figure 4.22. The column temperature was 210C and the eluent was 1-~hloronaphthalene.~~

5 References
1. 2. 3. 4. 5. 6. 7. 8. 9. 10.
11. 12. 13. 14. 15. 16. 17. 18. 19.

L.R. Snyder, J . Chromatogr. Sci., 1978, 16, 223. D.J. Pietrzyk and C.-H. Chu, Anal. Chem., 1977,49,860. T. Hanai, K.C. Tran, and J. Hubert, J . Chromatogr., 1982,239, 385. T. Hanai and J. Hubert, HRC & CC, 1984,7,524. Y. Arai, M. Hirukawa, and T. Hanai, Nihon Kagaku Kaishi, 1986,969. From the Mitsubishi Kasei catalogue. From the Tosoh catalogue. From the Asahipak catalogue (Showa Denkko). G. Schill, R. Modin, and B.-A. Person, Acta Pharm. Seuc., 1965,2, 119. B.L. Karger, S.C. Su, S. Maichese, and B.-A. Person, J. Chromatogr. Sci., 1974,12, 678. M. Hirukawa, M. Maeda, A. Tsuji, and T. Hanai, J. Chromatogr., 1990,507,95. S . Eksborg, P.-0. Lagerstrom, R. Modin, and G. Schill, J. Chromatogr., 1993,83,99. E. Eksborg and G. Schill, Anal. Chem., 1973,45,2029. M. Denkert, L. Hackzell, G. Schill, and E. Sjogren, J . Chromatogr., 1981,218, 31. L.E. Vera-Ada, M. Caude, and R. Rosset, Analusis, 1982, 10,43. K. Sugden and G.B. Loscombe, J . Chromatogr., 1978,149,377. B.A. Bildingmeyer and F.V. Warren, Anal. Chem., 1984,54,2351 . S.J. Valenty and P.E. Behnken, Anal. Chem., 1978,50, 834. B. Franssod, K.-G. Wahlund, I.M. Johansson, and G. Schill, J . Chrornatogr., 1976, 125, 327.

94

Chapter 4

20. J.F. Lawrence, F.E. Lancaster, and H.B. Conacher, J. Chromatogr., 1981,210,168. 21. M.G.M. de Ryter, R. Cronnelly, and N. Castagnoli, Jr., J. Chromatogr., 1980, 183, 193. 22. W. Voelter, K. Zech, P. Arnold, and G. Ludwig, J. Chromatogr., 1980,199, 345. 23. D.N. Buchanan, F.R. Fucek, and E.F. Domino, J. Chromatogr., 1980,181,329. 24. R.G. Peterson, B.H. Rumack, J.B. Sullivan, and A. Makowski, J. Chromatogr., 1980,188,420. 25. M. Ghias-Ud-Din, E.B. Olson, Jr., and J. Rankin, J . Chromatogr., 1980, 192,463. 26. K.T. Muir, J.H.G. Jonkman, D.-S. Tang, M. Kunitani, and S. Riegelman, J. Chromatogr., 1980,221,85. 27. M. Schoneshofer and A. Fenner, J. Chromatogr., 1981,224,472. 28. C.M. Riley, L.A. Sternson, and A.J. Repta, J. Chromatogr., 1983,276,93. 29. M. Kneczke, J. Chromatogr., 1980,198,529. 30. M. Alvinerie, P. Galtier, and G. Escoula, J. Chromatogr., 1981,223,445. 31. N. Kurosawa, S. Morishima, E. Owada, and K. Ito, J. Chromatogr., 1984,305,485. 32. D. Beales, R. Finch, and A.E. McLean, J. Cllromatogr., 1981 226,498. 33. P.B. Baker and T.A. Gough, J . Chromatogr. Sci., 1981,19,483. 34. J. Wagner, M. Palpeyman, and M. Zraika, J . Chromatogr., 1979,164,41. 35. R.V. Smith, J.C. Glade, and D.W. Humphrey, J . Chromatogr., 1979,172,520. 36. K. Makino, H. Ozaki, T. Matsumoto, H. Imaishi, and T. Takeuchi, J. Chromatogr., 1987,400,271. 37. Y. Kobayashi, H. Kubo, and T. Kinoshita, J. Chromatogr., 1987,400, 113. 38. G. Inchauspe, P. Delrieu, P. Dupin, M. Laurent, and D. Samain, J. Chromatogr., 1987,404, 53. 39. I. de Miguel, E. Puech-Costes, and D. Samain, J. Chromatogr., 1987,407, 109. 40. J.J. Lauff, J. Chromatogr., 1987,417,99. 41. K. Lindgren, J. Chromatogr., 1987,413, 351. 42. G. Lam, R.M. Williams, and C.C. Whitney, J. Chromalogr., 1987,413,309. 43. D.L. Theis, J. Chromatogr., 1987,402,335. 44. A. Sokolowski and K.-G. Wahlund, J. Chromatogr., 1980, 189,299. 45. Y. Yoshitoku, M. Moriyasu, and Y. Hashimoto, Chem. Lett., 1982, 193-194. 46. P.G. Rigas and D.J. Pietrzyk, Anal. Chem., 1987,59, 1388. 47. B.A. Bildingmeyer, C.T. Santasania, and F.V. Warren, Jr., Anal. Chem., 1987, 59, 1843. 48 L. Gagliardi, G. Cavazzutti, A. Amato, A. Basili, and D. Tonelli, J. Chromatogr., 1987,394,345. 49 M.L. Abaugh, 1,s.Krull, and W.R. Lacourse, J. Chromatogr., 1987,387, 301. J.H. McB. Miller, C. Pascal, and M. Tissieres, J . Chromatogr., 1987, 392, 361. 50. . J.A. De Shutter, W. Van Den Bossche, and P. De Moerloose, J. Chromatogr., 1987, 51. . 391, 303. 52 H. Lentzen and R. Simon, J . Chromatogr., 1987,389,444. 53 E. Naline, B. Flouvat, C. Advenier, and M. Pays, J . Chromatogr., 1987,419, 177. # 54. J.H. Kennedy and B.A. Olsen, J. Chromatogr., 1987,389, 369. 55 M. Dreux, M. Lafosse, P. Agbo-Hazoume, B. Chaabane-Doumandji, M. Gibert, and Y. Levi, J. Chromatogr., 1986,354, 116. 56 S.P. Sood, L.E. Sartori, D.P. Wittmer, and W.G. Haney, Anal. Chem., 1976,48,796. 57 K.-G. Wahlund and A. Sokolowski, J. Chromatogr., 1978, 151, 299. 58. N.D. Brown and H.K. Sleeman, J. Chromatogr., 1977, 138,449. 59. D.P. Wittmer, N.O. Nuessle, and W.G. Haney, Jr., Anal. Chem., 1975,47, 1422. 60. E. Fitzgerald, Anal. Chem., 1976,48, 1734. 61. J.H. Knox and J. Jurand, J. Chromatogr., 1975, 110, 103. 62. J.C. Kraak, K.M. Jonker, and J.F.K. Huber, J . Chromatogr., 1977,142,671. 63. K.-G. Wahlund and U. Lund, J. Chromatogr., 1976,122,269. 64. J.H. Knox and G.R. Laird, J . Chromatogr., 1976, 122, 17. 65. K.-G. Wahlund, J. Chromatogr., 1973, 115,411.

Selection of the Eluent

66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76.

95

J. Mitchell and C.J. Coscia, J. Chromatogr., 1978, 145, 295. W. Funasaka, T. Hanai, and K. Fujimura, J. Chromatogr. Sci., 1974,12,517. T. Hanai and K. Fujimura, J. Chromatogr. Sci., 1976, 14, 140. R. Neher, in Thin-layer chromatography, ed. G.B. Marini-Bettola, Elsevier, Amsterdam, 1964, p. 75. L.R. Snyder, J. Chromatogr., 1974,92,223. S. Hara, J. Chromatogr., 1977, 137, 199. S. Hara, Y. Fujii, M. Hirasawa, and S. Miyamoto, J. Chromatogr., 1978,149, 143. S. Hara and A. Ohsawa, J . Chromatogr., 1980,200,85. H. Hyakutake and T. Hanai, J . Chromatogr., 1975,108,385. From Shodex data. From Senshukagaku data.

Bibliography of additional references on theoretical approaches in ion-pair liquid chromatography

A. Bartha, H.A.H. Billiet, L.D. Galan, and G. Vigh, Studies in reversed-phase ion-pair chromatography 111. The effect of counter-ion concentration, J. Chromatogr., 1984,
291, 91.

A. Bartha, G . Vigh, H. Billiet, and L. de Galan, Effect of the type of ion-pairing reagent in reversed-phase ion-pair chromatography, Chromatographia, 1985,10, 587. B.A. Bidlingmeyer, S.N. Deming, W.P. Price, Jr., B. Sachok, and M. Petrusek, Retention mechanism for reversed-phase ion-pair liquid chromatography, J. Chromatogr., 1979,186,419. J.B. Green, Liquid chromatography on silica using mobile phases containing tetraalkylammonium hydroxides, J. Chromatogr., 1986,358, 53. M.T.W. Hearn (ed.), Ion-pair Chromatography, Marcel Dekker, New York, 1985. S.O. Jansson, Effect of counter-ions in ion-pair liquid chromatography of hydrophobic amines on non-polar bonded phases, J . Liq. Chromatogr., 1982,5,677. J.C. Kraak, H.H. van Rooij, and J.L.G. Thus, Reversed-phase ion-pair systems for the prediction of n-octanollwater partition coefficients of basic compounds by highperformance liquid chromatography, J. Chromatogr., 1986,352,455. W.R. Melander and Cs. Horvath, Mechanistic study of ion-pair reversed-phase chromatography, J. Chromatogr., 1980, 201, 21 1. M. Puttemans, L. Dryon, and D.L. Massart, Extraction of organic acids by ion-pair formation with tri-N-octylamine, Anal. Chim. Acta, 1984, 161,221. W. Santi, J.M. Huen, and R.W. Frei, High-speed ion-pair partition chromatography in pharmaceutical analysis, J . Chromatogr., 1975,115,423. G. Schill, Selective extraction of organic compounds as ion-pairs and adducts, Talanta,
1975,22, 1017.

M. Tang and S.N. Deming, Interfacial tension effects of nonionic surfactants in reversed-phase liquid chromatography, Anal. Chem., 1983,55,425. O.A.G.J. van der Houwen, R.H.A. Sore], A. Hulshoff, J. Teeuwsen, and A.W.M. Indemans, Ion-exchange phenomena and concomitant pH shifts on the equilibration of reversed-phase stationary phase materials with ion-pairing reagents, J. Chromatogr., 1981, 209, 393.

CHAPTER 5

Separation Based on an Improved Column Eficiency

High-speed separation can be achieved by reducing the column length, column internal diameter, the size of the stationary phase particles, and by increasing the eluent flow velocity, as explained in Chapter 1. In addition, the development and optimization of pumps, detectors, injectors, and the recording systems contribute to improved separations. The influence of diffusion theory on the experimental separation should also be considered. If the unresolved chromatogram shown in Figure 5.1A has been obtained for two compounds, two methods can be applied to achieve a complete separation. Figure 5.1B demonstrates that achieved by improving selectivity (relative retentions), and Figure 5.1C demonstrates a separation achieved by improving the column efficiency. The resolution, R,, is used to describe the separation conditions of two compounds. The R, value is given by the ratio of the distance between

elution volume

-+

Figure 5.1 Improvement of separation. A, unresolved peaks; B, increase in peak separation; C , improved peak eficiency.

96

Separation Based on an Improved Column Efficiency

97
wb,

the maxima of two peaks, A ~ R and the mean of the peak widths, , base:

at their

where tR1and tR2 are the retention times of the peaks, and wbl and wb2 are the base widths of the two peaks. Theoretically, when R, is 1, the two peaks are 98% separated. An R, value > 1 indicates a good separation and an R, value c 1 indicates a poor separation. In Figure 5.2, the resolution between peaks 3 and 4 was 1.1. The resolution between peaks 4 and 5 was 1.5, and that between peaks 5 and 6 was 2.4. The resolutions do not exactly match with their respective peak shapes because of their different intensities. If two symmetric peaks have a separation that results in a resolution of R, = 1, when the peaks have equal
1
5

I 0

4 1

8 min 2 mL
-~

peak
plate number N Rs
Figure 5.2

3
3,364

4
2.116

5
3.330

6
1.129

1.1

1.5

2.4

Normal phase separation o phthalates. Column, Yanapak C N , f 25 cm x 2 mm id.; eluent, n-hexane-n-butanol (200.1); flow rate, 0.25 ml min-'; pressure, 3 MPa. Peaks: 1, lauryl phthalate; 2, heptyl phthalate; 3, butyl phthalate; 4, propyl phthalate; 5 , ethyl phthalate; and 6 , methylphthalate.

98

Chapter 5

areas (Figure 5.3A), the resolution will deteriorate if there is a change in the ratio of the areas of the peaks, even if the retention times remain constant. An example is shown in Figure 5.3B;as the ratio increases the experimental resolution decreases. In designing a procedure it is therefore desirable to obtain equal peak areas for overlapping peaks. When the wavelength of the isosbestic point of the spectra of two compounds is used for photometric detection, the peak areas of a 1:l mixture will be equal. Alternatively, the selection of the wavelength can be used to generate equal peak areas even if the mixture ratio is very different. When pure standard chemicals are used, their spectra should first be measured, and the most suitable wavelength selected to obtain good resolution and the correct peak area ratio. In practice, the selectivity depends on the structure of the analytes and the mixture ratio, and it is therefore often difficult to obtain an appropriate relationship between the R, value and the mixture ratio. When one large peak is observed on a chromatogram, the purity of the peak should be studied by using a longer column or changing the detector wavelength. Conversely, poor resolution on a chromatogram can be improved by appropriate selection of the detector wavelength. An improvement in the resolution can also be achieved by increasing the separation of the peak (AtR) or narrowing the peak width (increased N ) . An increase in AtR depends on the selection of stationary phase materials and eluent; a reduction of the peak width depends on improving the column
(A)
volume ratio
1:l 2:l 4:1 8:1 32:l 128:l 16:l

theoretical
resolution 1.0 1.25

(B)

experimental
volume resolution ratio
1:l 1.46

0.8

h
Jisi

2:l 4:l

1.46 1.32

8:l

0.84

Figure 5 3 Resolution related to peak area ratio. (A) theoretical and (B) experimental . changes in resolution with volume ratio. Column, 10 pm octadecyl-bonded silica gel, 15 cm x 4.6 mm id.; eluent, 90% aqueous acetonitri1e;jow rate, 1 ml min-; detection, UV 254 nm. Sample: diflerent ratios of 10 pl benzene and toluene in 10 ml ethanol.

Separation Based on an Improved Column Efficiency

99

efficiency. The enhancement of high-performance liquid chromatography depends mainly on improving the column efficiency.

1 Improving Separation by Changing the Selectivity

The selectivity of a separation is based on differences in the physicochemical interactions of the analytes with the stationary and mobile phases. This difference is measured as AtR on a chromatogram. The retention volume of each component, V,, is given by:

where V Mis the volume of mobile phase inside the column, Vs is the volume of the stationary phase in reversed-phase chromatography, and K is the distribution constant of the analyte between the mobile and stationary phases. In normal-phase liquid chromatography, V , is related to the surface area of the stationary phase material, and in ion-exchange liquid chromatography to the ion-exchange capacity. Differences in K values between analytes in one system reflect differences in the properties of the compounds. In any one column the distribution between the stationary and mobile phases is described by the mass distribution coefficient. To change values of AK between analytes, either the phases must be changed (changes in stationary and mobile phase selectivity, see earlier) or the temperature must be changed. Mass distribution constant = amount of solute in stationary phase amount of solute in mobile phase

where Vs and V Mare constant for all analytes in a particular column. This value is equal to the retention factor k . The stronger is the interaction between the sample molecule and the stationary phase the larger the retention of solute. Thus the larger the difference in K between analytes (AK), the better will be their separation. To improve the separation, the difference between the distribution constants ( K values) of compounds 1 and 2 must be increased, or the volume of the stationary phase ( V s )must be increased.

To increase Vs, the chromatographer can increase the surface area of the stationary phase materials in normal-phase liquid chromatography, increase the stationary phase volume in reversed-phase or partition liquid chromatography, or increase the ion-exchange capacity in ion-exchange liquid chromatography. In general, if the internal diameter of a column is constant, the retention time

100

Chapter 5

difference (AtR) is related to the column length (L). Therefore, doubling the column length doubles the difference A t R . However, the resolution of the is analyte (R,) only related to ,/L due to the diffusion of the samples inside the column. There is no molecular interaction in size-exclusion liquid chromatography, and therefore the resolution can only be improved by increasing the column length.

2 Improving Separation by Increasing the Column Efficiency

The peak broadening due to diffusion is related to the square root of L as follows:

where H i s the column efficiency, i.e. the height equivalent to a theoretical plate (HETP). A smaller H value gives better resolution. From Equation 5.3:

If a peak diffuses as a theoretical Gaussian distribution, peak width is 40 standard deviation, from Equation 5.4: H c 02/L c

wb

is equal to 40. One

(5-5)

The separation factor a of two compounds is given from Figure 3.4:

a=
tR2 tR1

- tRO - v R 2 - VO - - K2 - tRO V R l - VO K1

The number of theoretical plates is given by:

N = 16 - - 16

(2)2- (2)2=4 5 ( 3 2 = (y 5.5

The effective number of theoretical plates is the number of usable theoretical plates and is calculated after elimination of the void volume:

If the peak widths of peaks 1 and 2 are assumed to be equal, the following equation is derived from Equation 5.1:

Separation Based on an Improved Column Efficiency

101

From Equation 5.7,

*
wb2
=

=0 . 2 5 a

(5.10)

From Equation 5.2, ( V R ~ VM) = K Vs, ( VRI- VM) = K1 Vs 2 Hence:

From Equation 5.6, a

K 2 / K l ,therefore

The mass distribution coefficient k = K,V,/VM

(VR - VM)/VM, and

Therefore, from Equation 5.9:

R ~ = - x - a - 1 ~ k x m = ~ ( - 1 - a - 1- & ) f i 1 - -& 4 a

(5.11)

Combination of Equations 5.8 and 5.1 1 gives: (5.12) The relationships between R,,a, and N are given in Figure 5.4. When a is small, a large N is required for a good separation. However, the use of a longer column should be avoided for fast analysis even if a longer column provides excellent plate numbers. An improvement in the value of a can be estimated for a good separation. A column having a smaller HETP value is a good column because diffusion inside the column is small, resulting in better separation. The HETP value is given by the Van Deemter equation, which describes the peak broadening of packed columns through which a non-compressible solvent is moving.

102

Chapter 5

10 5

0,

5
E

104

1.01

1.10

1.20

1.30

a
Figure 5.4 Relationship between separation factor a, resolution R,, and plate number N . Line: 1, R, = 1.5; 2, R, = 1.2.5; 3 , R, = 1.0; and4, R, = 0.8.

H = H F + H L + Hs + H M .The H value can be expressed as the sum of (a) eddy diffusion ( H F ) , (b) longitudinal diffusion (HL), and (c) resistance to mass respectively). transport in stationary and mobile phases (Hs and H M ,

Reducing the Eddy Diffusion Term

Eddy diffusion is a result of the presence of particles of stationary phase material in a column, and depends on the stationary phase conditions, shape of the column, and the structure of the stationary phase material. The influence of the stationary phase material can be divided into the particle size (d,,), the shape of the particles, and the porosity of the particles. The standard deviation for peak broadening due to the particles is described by: oF2= 2AdpL, and therefore from Equation 5.5:
HF
=

2Adp

(5.13)

1 depends on the irregularity of the particles (particle shape) and on the column
material (steel gives more disturbance than glass); the relative effect will always increase as the column diameter decreases. The above-mentioned effects have led to an optimum internal diameter of 2-5 mm for LC columns. In a column with this internal diameter, which is uniformly packed with small spherical particles, the eddy diffusion will be limited. The A value is about 1 for spherical particles.

Separation Based on an Improved Column Efficiency

103

Reducing the Effect of Longitudinal Diffusion

Molecules tend to diffuse randomly, in no particular direction, within any fluid, independently of the flow rate of the mobile phase. Their diffusion rate is determined by the type of molecule, the nature of the mobile phase, and the temperature, and is expressed quantitatively by their diffusion constants. Molecular diffusion is related to time, as shown by the Einstein equation:
o2 = 2Dt

(5.14)

where D is the diffusion coefficient. The standard deviation due to longitudinal diffusion is oL, and the time during which diffusion can occur is t, = L/u; therefore, oL2 2 y ~ , t , = (2 7 4 . &)/# , and from Equation 5.5, H = 0 2 / ~ =

where y is the factor of anti-diffusion, and is less than 1; DM is the diffusion constant of sample molecules in the mobile phase; and u is the linear velocity of the mobile phase. The term HL is important in gas chromatography because D M is high. The DM value in liquids is ca. low5, and the flow rate is greater than the diffusion speed; therefore, HL is not as important. The HL value is reduced by an increase in the viscosity of the solvent or by a decrease in the temperature. Longitudinal diffusion can thus be reduced by decreasing the diffusion coefficient and increasing the flow rate; however, these two actions are counter-effective in liquid chromatography because of the mass transport term.

Reducing the Effect of Resistance to Mass Transport in the Stationary and Mobile Phases
This dispersion effect results from the mass transfer of the analyte between the stationary and mobile phases, and is an important phenomenon in highperformance liquid chromatography. The injected analyte molecules are first present in the mobile phase. They then transfer back and forth from the stationary phase in order to interact. This process is repeated along the column, from the inlet to outlet. Such a phenomenon occurs in all kinds of chromatographies: normal-phase, reversed-phase, partition and ion-exchange liquid chromatography. (In size-exclusion liquid chromatography, such movement of sample molecules occurs as a diffusion process between the inside and outside of the stationary phase material.) Under ideal conditions, the sample molecules move along the column at the same speed in both the mobile and stationary phases (dotted lines in Figure 5.5). In practice, the movement of the mobile phase drags the analyte along faster than their movement through the stationary phase because equilibration at the interface is restricted by the diffusion rates within the phases. Therefore peak broadening occurs according to the total diffusion profile, as shown in Figure 5.5. The total diffusion effect

104
Total diffusion

Chapter 5

Mobile

phase
Stationary
, -

phase

u - - . i d e a l diffusion Diffusion in stationary

, t,

; ,

Flow direction

Rhase

f Figure 5.5 Diffusion o analyte in mobile and stationary phases while moving along the column.

can be divided into diffusion in the stationary phase ( H s ) and that in the mobile phase (Ifhl). Within a certain time ( t a ) , the sample molecules in the mobile phase move a particular distance (Uta) at a linear velocity (u). During this period, the movement of the sample molecules in the stationary phase is zero. If the number of sample molecules transferred between the mobile and stationary phases is defined as n,, and the distance the centre of the band has moved is defined as dR; n, can be related to 2dR/uta. The time the analyte spends in the stationary phase is ?d. The ratio R of the moving speed of the band and the mobile phase is defined as ta/(ta + t d ) . The following equation is then obtained:

The distance (I,) that the band centre has moved is related to R and Utd:

The standard deviation caused by molecular diffusion is:

o2 =

12ns

(5.16)

Therefore, the standard deviation due to molecular diffusion inside the stationary phase (as) is:
as2 = 2qR(1 - R ) u L t d

(5.17)

where q is a parameter that depends on the type of stationary phase used, e.g. spherical, irregular, fiber or porous. The q value is 213 for pellicular-type stationary phase materials and 1/30 for porous-type stationary phase materials. The average residence time ( t d ) of the sample molecules in the stationary phase is related to the thickness of the stationary phase (ds) and the diffusion constant ( D s ) of the sample molecules in the stationary phase. Thus 0 = 2Dt (Equation 5.14) and ?d = ds2/2Ds is obtained.

Separation Based on an Improved Column Efficiency

105
=

R = VM( + KVs) and D VM M Equation 5.17 is converted into:

KVs/ V,, therefore R

1/(1 + DM). Then,

Therefore, the band broadening due to diffusion of sample molecules in the stationary phase (Hs)is given by:

From this equation, the Hs value is small in a high-efficiency column. Such a column has a stationary phase with a thin film or small ds value. That is, a highperformance separation can be achieved by using a stationary phase in which the depth into which the sample molecules can move is diminished. This means that a thinner or shallower stationary phase gives higher performance in reversed-phase liquid chromatography, and a smaller particle size gives higher performance in ion-exchange and size-exclusion liquid chromatography. However, speeding up the diffusion counteracts lowering the flow rate. For diffusion in the mobile phase, we can consider the eddy diffusion (HF) of Equation 5.13, longitudinal diffusion (HL)of Equation 5.15, and the diffusion toward the wall from the centre of a column (or space between particles) (HD). The time taken for the sample molecules to move from the column side wall or particle surface is given by Equation 5.14, where the distance is given as wadp:

where w, is the ratio between the distance moved to the wall and the particle diameter (dp). When cob is defined as the ratio between limited partial linear velocity and the average linear velocity, the distance for sample molecules movement ( I ) is opt. Therefore, the standard deviation, cr, is:

where w

oa2wp2/2. Then the diffusion effect due to the mobile phase, HD, is: (5.19)

where co is the column constant, which depends on the shape of the column and the packing conditions. As measured, the Hs term means that a smaller particle size reduces the distance between particles, and thus the spreading due to diffusion of sample molecules is minimized. From Equations 5.18 and 5.19, a decrease in H value is achieved by increasing the diffusion speed (elevating the column temperature,

106

Chapter 5

using a low-density, low-viscosity solvent), decreasing the flow rate, reducing the thickness of the stationary phase, and decreasing the particle size. The overall effects of peak diffusion can be summarized as follows:

Overall, the most effective factor in Equation 5.20 is the particle size. The smaller the particle size, the higher the column efficiency. Equations 5.13, 5.15, and 5.18 are depicted in Figure 5.6 against flow velocity as A, B, and C , respectively. The band spreading is thus influenced by Equation 5.15 at a low flow rate. The band spreading is influenced by Equations 5.18 and 5.19 at a high flow rate. For gas chromatography curve D is obtained. In liquid chromatography, the diffusion rates are slower than that in gas chromatography, and the values of D and Ds are very small; therefore, the M minimum H value is obtained at a low flow rate, as shown by curve E in Figure 5.6. The value of H increases slowly at higher flow rates in liquid chromatography. An experimental result is shown in Figure 5.7. The HETP was minimal at a certain flow rate, and the measured optimum value was less than 10 pm for this column. The optimum flow rate was about 0.9 ml min- corresponding to a linear flow velocity of about 55 mm min-'. Experimentally, high-performance separations can be performed at lower flow rates using small particle-size stationary phase materials as shown in Figure 1.1. A fast analysis can be achieved at a higher flow rate when the H

',

n c w I

linear velocity

cdmin

Figure 5.6

Van Deemter curves relating H to linearflow velocity. A, eddy difusion term; B, molecular difusion term; C , resistance to mass transfer term; D, summation for gas chromatography; and E, summation for liquid chromatography.

Separation Based on an Improved Column Efficiency

flow rate 1.0 mL I min

107

benzene toluene ethylbenzene propylbenzene butylbenzene pentylbenzene

0.86

k = 1.18 k
k

= 1.55 =
2.18

+
A

k = 3.18

k = 4.34

N = 100,000 / m

20

40

60

80

100

Linear flow velocity

mdmin

f Figure 5.7 HETP o compounds with diflerent k values related to linear flow velocity. Column: 5 pm ODs-bonded silica gel, 15 cm x 4.6 mm i.d.; eluent: 80% aqueous acetonitrile.

value is smaller; however, Equations 5.18 and 5.19 indicate that a high flow rate operation has negative effect. Curve E in Figure 5.6 indicates that a high flow rate operation does not significantly increase the H value in liquid chromatography. This means that a fast, efficient analysis can be performed by using a short column packed with small particle size stationary phase materials. The pressure drop, Ap, inside a column is given by Equation 5.21:

where tj is the viscosity of the eluent, L is the column length, u is the linear flow velocity,fis the sum of the void volume of a column, and K O is the parameter indicating the column permeability. When the void volume between the stationary phase materials is E , K O is given by Equation 5.22:
KO=-X-

dp2 E~ 180 (1 - E ) 2

(5.22)

From

tR =

L(DM + l)/u and Equation 5.21, the following equation is obtained:

(5.23)
Thus when q , f , and L are constant, shortening the analysis time can be achieved by increasing the pressure drop and the K O value. Increasing K O can be done by decreasing E and/or increasing the particle size, dp, of the stationary phase material. Increasing the particle size decreases the resolution, R,;therefore, small particle-size stationary phase materials must be densely packed. As a result, a high-pressure operation is unavoidable.

108

Chapter 5

In conclusion, the smaller is the particle size the better the separation; however, a high pressure drop requires the optimum design of columns and instruments. A column well packed with spherical small particles has a high plate number, and is suitable for the separation of a homologous series of compounds. However, the separation of isomers requires a high selectivity of stationary and mobile phases as described in Chapters 3 and 4.

3 Bibliography
See Chapter 1 for details.

CHAPTER 6

Influence o Physical Chemistry f on Separations in Liquid Chromatography

1 Quantitative Structure-Retention Relationships in Reversed-phase Liquid Chromatography
Various methods have been employed for the prediction of retention times in reversed-phase liquid chromatography. As the alkyl chain length of fatty acids increases, their solubility in water becomes poorer and that in n-hexane becomes higher. Such a phenomenon is easily understood from a comparison of the properties of acetic acid and butter fat. In reversed-phase liquid chromatography, the retention time of fatty acids is related to the number of methylene units, i.e. the chain length. This relationship is a basic concept used to study retention times in reversed-phase liquid chromatography. In early work the number of carbon atoms of the aliphatic chain was related to the retention of alcohol^'^ and Even the polymerization of very polar compounds, such as amino acids and monosaccharides, makes them less soluble in water than their monomers. This means that polymerized molecules become more hydrophobic and their retention times become longer in reversed-phase liquid chromatography. Therefore, the retention time of a homologous series of compounds is easily predicted from the number of monomer units, and sometimes by the carbon numbers. It can also be related to molecular size, van der Waals volume, and surface area. A quantitative analysis of the structure-retention relationship can be derived by using the relative solubility of solutes in water. One parameter is the partition coefficient, log P, of the analyte measured as the octanol-water partition distribution. In early work, reversed-phase liquid chromatography was used to measure log P values for drug design. Log P values were later used to predict the retention times in reversed-phase liquid chromatography.The calculation of the molecular properties can be performed with the aid of computational chemical calculations. In this chapter, examples of these quantitative structure-retention relationships are described.
109

110

Chapter 6

Prediction of Retention Times from log P in Reversed-phase Liquid Chromatography Prediction of Partition Coeficient (log P )
The measurement of the solubility of drugs in polar and non-polar media is very important in the pharmaceutical field. One method proposed to describe this solubility is the partition coefficient between octanol and water. The mathematical calculation of an octanol-water partition coefficient from values for functional groups was first proposed by Hansch et al. as Hansch's n constants,' and was later developed by Rekker as hydrophobic fragmental constants (log P).2 This method was further improved by the use of molecular connectivities.17 The prediction of logP values can be performed by either a computer program or by manual calculation. For example, approximate partition coefficients (log P)have been calculated by Rekker's method:

A part of his fragmental constant tables is given in Table 6.1. (The predicted log P values are summarized in Table 6.4.) Attempts have also been made to determine octanol-water partition coefficients empirically using liquid chromatography by comparison of retention times with those of compounds of known log P.3-'2

Correlation between log P and log k Values

It seemed possible to correlate the elution order in reversed-phase liquid chromatography with the octanol-water partition coefficient, log P. For example, the partition coefficients calculated by Rekker's method showed a

Table 6.1 Rekker 's hydrophobic fragment constants

Fragment

A log P

Fragment

A log P
-0.954 - 0.093
- 1.292

Fragment

A log P
- 1.155 -0.581

1.886 1.688 1.431 0.702 0.530 0.235 0.36 0.15 0.935 0.73 0.175

(a1)COOH (ar)COOH (al)C00 (ar)COO (a1)CO (ar)CO (a1P (am (a1)OH (ar)OH (ar)COH

(al)OCH2COOH (ar)OCH2COOH

- 0.43 1 - 1.703

-0.842 - 1.581 -0.433 - 1.491 -0.343 -0.38

ar: substituent on aromatic compound, al: substituent on aliphatic compound; hydrophobic fragment constants in octanol-water from Rekker2.

Influence of Physical Chemistry on Separations in Liquid Chromatography

111

linear relation to log k values measured in reversed-phase liquid chromatography: l 8 logk=y x logP+m (6.2)

where y and rn are constants in a given system. An example is given in Figure 6.1. The values of log P obtained from Equation 6.1 using the measured log k of 58 compounds, including alkanols, alkylphenones, alkyl benzoates, polycyclic aromatic hydrocarbons, halogenated benzenes, etc., were related to the log P values calculated according to Rekkers method. The average difference was within 0.16 log P unit. The correlation coefficient was 0.975. However, the peak for aniline showed tailing. An organic modifier effect on logP was described, and a method for the qualitative analysis of free fatty acids was proposed. 39 8920

Prediction of Retention Timesfrom log P

The prediction of retention times in a given eluent from log P has been proposed for aromatic hydrocarbons. The log k values of phenols2 and nitrogencontaining compounds22 were also related to their logP, and the calculated log P was used for the qualitative analysis of urinary aromatic acids, i.e. for the identification of metabolites in urine from the differences of l o g P in reversedphase liquid chromatography.23724 A good correlation was obtained in 2&80% acetonitrile-water mixtures. The standard non-ionic compounds used to evaluate the columns were 2-hydroxyacetophenone, coumarin, acetophenone, indole, propiophenone, butyrophenone, isopropyl benzoate, butyl benzoate, and isopentyl benzoate. The plotted lines for the linear relationship measured in five different proportions

-2

logP by Rekkers Figure 6.1

Relationship of Rekkers log P values to log k values on octadecyl-bonded silica gel column at 30C. Column, Chromosorb LC7; eluent, 60% aqueous acetonitrile. Number beside symbols: see Table 6.4.

112

Chapter 6

of acetonitrile-water mixtures merged at a single point. These slopes ( y ) have a fourth-degree polynomial relation with the acetonitrile concentration ( x ) (Equation 6.3, where ai is a contant):
y=

Ca;x xi
i=O

i=n

n = 2-8

Further, an optimization of the organic modifier concentration for the separation of known compounds was proposed. When the value of the actual column plate number, the resolution, and log P of pairs of compounds a and b were known, the percentage concentration (x)required for their separation was calculated from Equations 6.4 to 6.6:'9720

R, = [(a- l)/a]O.25&

(6-4)

The difference between the logk values measured and those predicted by Equations 6.1 and 6.2 from log P values calculated by Rekker's hydrophobic fragmental constant was within 5%. An example of logk values obtained experimentally compared with those predicted from Rekker's log P values is shown in Figure 6.2, where the predicted and measured retention factors of aromatic acids are shown. The correlation coefficient was 0.993 (n = 32), and the slope was 1.014. The precision is very high.
50
L
c ,

0
0

rc

(P

40

0 .-

c ,

30

Q) c , Q)

3 0 .U

20

c ,

! !

10

0 0

10

20

30

40

!D

Measured retention factor

Figure 6.2 Comparison of measured and predicted retention factors of aromatic acids from Rekker's log P values. Column, 10 pm polystyrene gel, Hitachi 3011, 15 cm x 4.6 mm i.d.; eluent, 30% aqueous acetonitrile containing 50 mM phosphoric acid at 55 "C. Numbers beside symbols: see Table 6.4.

Influence o Physical Chemistry on Separations in Liquid Chromatography f

113

Prediction of Retention Time of Ionic Compoundsfrom log P and

PK,
The retention times of neutral non-ionized compounds can be predicted by the above calculation method, but those of ionized compounds are also very important. The inclusion of the dissociation constant in the calculation made it possible to predict the retention times of ionized aromatic The dissociation constant was calculated by the method proposed by Perrin et When a compound is ionized, the retention becomes shorter than that of the molecular form, and the retention factor of the ionized form is obtained from Equation 6.7:

where k is the retention factor of the non-ionized acid and can be obtained m from log P as above and ki is the retention factor of the 100% ionized acid. The ki value cannot be predicted mathematically at present, but the value is close to zero in many cases. Equation 6.7 was further modified to improve the precision in low pH solutions where the strong acid used for pH control reduced the retention of weak acids by excluding them from the hydrophobic surface of the stationary phase.26 A modified version of Equation 6.7 is given in the following form (a slight modification of constant A improved the precision):

An example of the relationship between the predicted and observed log k values at pH 4.00 is shown in Figure 6.3. The agreement between the observed and predicted k values of aromatic acids was within 10%. The correlation coefficient was 0.954 ( n = 32). An error of greater than 10% for 3-hydroxy-2-naphthoic acid and 2-hydroxybenzoic acid was attributed mainly to an error in their Ka values.25The partition coefficient, logP, and dissociation constant, pKa, of analytes can be obtained by simple calculations and by computational chemical calculations, and thus the retention time can be predicted in reversed-phase liquid chromatography.

Calculation of Dissociation Constant p K , from Hammets Equation

The dissociation constant of an analyte can be calculated mathematically from Hammets equation.27 The organic solvent effect on the pKa has also been examined:26

114

Chapter 6

20

measured retention factor at pH 4.00 Figure 6.3 Comparison of measured and predicted retention factors of aromatic acids from Rekker's log P values. Column, 5 pm polystyrene gel, Hitachi 3013, 15 cm x 4.6 mm i.d.; eluent, 20% aqueous acetonitrile containing 50 mM sodium phosphate at 55 "C. Numbers beside symbols: see Table 6.4.

where A and B are constants for individual groups of compounds, and Q is Hammet's Q constant. However, it is difficult to predict the pK, values of orthosubstituted compounds. Examples of pK,s from modified Hammet's equations are:25 Benzoic acids: 2-Hydroxybenzoic acids: Phenylacetic acids: Cinnamic acids: Mandelic acids: 4.20-1 .OO C Q 4.20-1.13 C Q 4.30-0.35 Q 4.38-0.66 C CT 3.38-0.454 C CT

The original equations for phenolic and nitrogen-containing compounds are:27 Phenolic Compounds: Anilines Pyridines (1): Pyridines (2): Quinolines (2-substituted) Quinolines (8-substituted) a-Naph t h ylamines: P-Naph t h y lamines: Benzylamines (ring substituted): N-Substituted anilines 1-Aminoanthracene: Quinoline: where
Q =

9.92-2.23 C G 4.58-2.28 Q 5.25-5.90 C CT 5.39-5.70 C Q 5.12-9.04 Y 0 4.64-3.1 1 0 3.85-2.81 C Q 4.29-2.8 1 C Q 9.39-1.05 Q 5.06-3.46 C CT aniline + 0.17 pyridine (1) + 0.06

Hammet's sigma constant. Examples of

constants are given in

Influence of Physical Chemistry on Separations in Liquid Chromatography

115

Table 6.2 Hammets substituent constants

Substituent
nmeta

Upara

Substituent

urnera

npara

Br

c 1
H I OH
From Ref. 27.

0.39 0.37 0.00 0.35 0.13

0.22 0.24 0.00 0.15 - 0.38

CH3 OCH3 NH2 COzH C2H5

-0.06 0.11 0.00 0.35 -0.07

-0.14 -0.28 -0.57 0.44 -0.15

Table 6.2. The calculated values are summarized in Table 6.4 (along with reference values).

2 Van der Waals Volume as the Basic Property

In a reversed-phase liquid chromatography system, if the hydrogen-bonding and Coulombic forces are negligible, the retention of molecules depends upon their size, and the presence of 71 electrons enhances the selectivity. The molecular size, i.e. the van der Waals volume, can be calculated by hand, by Bondis method,29 or by certain computer programs. The possibility for predicting the retention time of polycyclic aromatic hydrocarbons from a combination of the calculated van der Waals volume and the reference values of the delocalization energy has been d e m ~ n s t r a t e dThe~ .~ precision of a method for predicting the retention times of polycyclic aromatic hydrocarbons in reversed-phase liquid chromatography has been improved by a combination of the van der Waals volume calculated by the computer program MOPAC-BlogP and the kinetic energy calculated by molecular dynamic^.^' However, the precision for polychlorinated benzenes could not be improved by this method.31 The prediction of retention times of phenols was further improved by using the ionization potential calculated by the computational chemical program CACheTM.

Calculation of van der Waals Volume

Bondis calculation method is simple. The van der Waals volumes are the sum of the van der Waals volumes of fragments, as given in Table 6.3. The calculated van der Waals volumes are summarized in Table 6.4. However, these volumes are different from those calculated using the MOPAC-BlogP program, even though the correlation is excellent.

Prediction of Retention Times Based on van der Waals Volumes

The relationship between the log k and the van der Waals volumes of alkanes, alkylbenzenes, and polycyclic aromatic hydrocarbons measured in 70% aqueous acetonitrile on octadecyl-bonded silica gels3 is shown in Figure 6.4.

116

Chapter 6

Table 6.3 Bondi's group contributions to the van der Waals volume of hydrocarbons
Groups Alkane C CH CH2 CH3 CH4 Aromatic
/ \-(cond.)

VWV" 3.33 6.78 10.23 13.67 17.12

Groups

VWV" 6.96

k=C/

10.02 8.47 11.94 16.95 13.49

=CH=CH2
/

\C=CH2 \C=CH-

4.74 5.54 8.05 48.36 45.84

\-R
/

Acetylenic 8.05 rC-H EC-( diacet y lene)

') from Ref. 28.

\-H

Benzene Phenyl

11.55 7.82

" Van der Waals volume (cm3 mol-

1.5

4 0 6 2

:x 3 9
0
93

0 2SA
g7

O
1
Y

OA
0

A
0.5

I
-0.5 ! 25

3
Id

32

A A

75

125

175

van der Waals volume by Bondi Figure 6.4

cm3/mole

Relationship of log k values measured on ODs-bonded silica gel to Bondi's van der Waals volumes. Column, Develosil O D s , 15 cm x 4.6 mm i.d.; eluent, 70% aqueous acetonitrile at 30 "C. Numbers beside symbols: see Table 6.4: 0 , Polycyclic aromatic hydrocarbons; x , alkylbenzenes; 0, halogenated alkanols; and 0, alkanes. benzenes; A,

Table 6.4 Molecular properties of non-ionic compounds

VWV"

No. Chemical

Mr
SAb
VWVf

IF
Slog@ -AHe logp

-AHh

% ff 5
$
fi'

58.03 68.26 78.49 88.72 2.994 3.524 4.054 4.584

-

2.301 2.367 2.957 3.372 5.644

-

109.18

g
6.704
-

i r 3
<
2.280 3.208 1.484 1.926
0

129.64
-

43.86 73.96
-

3
5
-

g.
2 i

& 5
-

90.08 93.22 99.56 99.56 109.04 125.16 125.16

-

3.772 3.906 4.378 4.378 5.030 5.545 5.790

I

2.034 2.235 2.476 2.543 2.885 3.207 3.442

-

134.64
-

5 %
-

c
6.200

1 Pentane 2 Hexane 3 Heptane 4 Octane 5 Nonane 6 Decane 7 Undecane 8 Dodecane 9 Tridecane 10 Benzene 11 Naphthalene 12 Acenaphthylene 13 Acenaphthene 14 Biphenyl 15 Fluorene 16 Phenanthrene 17 Anthracene 18 Pyrene 19 Fluorancene 20 Chrysene 21 Tetracene 22 Benzo[a]anthracene 23 p-Terphenyl 24 Benzo[e]pyrene 25 Benzopyrene 26 Benzo[k]fluoranthene 27 Benzo[b]fluorancene 95.970 112.787 129.552 146.408 163.232 180.007 196.800 21 3.653 230.555 83.789 127.598 144.969 150.675 155.227 161.639 171.286 171.486 187.603 188.723 214.558 215.352 215.841 226.657 230.588 23 1.048 232.418 232.462 136.780 158.420 180.714 202.272 224.205 246.467 267.965 289.660 3 11.976 109.752 156.999 173.766 181.904 189.990 194.923 20 1.409 204.540 216.623 220.619 244.822 252.509 25 1.956 273.024 256.615 260.079 268.140 267.348 11.060 11.060 11.060 11.060 11.064 11.062 11.063 11.063 11.032 9.652 8.71 1 8.943 8.492 8.950 8.717 8.617 8.123 8.130 8.629 8.371 7.747 7.938 8.746 8.214 7.918 8.302 8.559 3.395 3.972 4.570 5.139 5.698 6.267 6.738 7.209 7.648 2.432 3.729 4.274 4.297 4.316 4.545 4.799 4.808 5.246 5.248 5.581 5.634 5.554 5.710 5.899 5.912 5.962 5.941 5175 6048 6919 7794 8668 9539 10411 11286 12164 3402 4897 5359 5668 6237 6145 65 19 6539 6847 699 1 820 1 8733 8470 9205 8595 8571 8669 8663

72.150 86.177 100.203 114.230 128.257 142.284 156.311 170.337 184.364 78.1 13 128.173 152.195 154.211 154.211 166.222 178.233 178.233 202.255 202.255 228.293 228.293 228.293 230.309 252.3 15 252.3 15 252.3 15 252.315

3.700

(continued) Z

''
0

F
5

c
w

Table 6.4 Continued

VWY"
Ip"
-

00

~~~

No. Chemical

M r

SAb
Slog@ VWVf

-AHe

1ogP

-AHh

150.76
-

6.960
-

59.51 69.74 77.02 79.97 90.20

-

2.588 3.1 18 3.525 3.648 4.178

-

1.658 1.794 1.906 2.053 2.331

-

110.66 120.89 131.12 141.35 151.58

-

5.238 5.758 6.297 3.453 3.655 1.75 2.28 2.81

2.797 3.200 3.673 6.828

3
q

28 Benzo[ghi]perylene 29 Dibenzo[ah]anthracene 30 Pentacene 3 1 Indeno[ 1,2,3-cd]pyrene 32 Toluene 33 Ethylbenzene 34 Isopropylbenzene 35 Propylbenzene 36 Butylbenzene 37 Pentylbenzene 38 Hexylbenzene 39 Heptylbenzene 40 Octylbenzene 41 Nonylbenzene 42 Decylbenzene 43 Acetophenone 44 Propiophenone 45 Butyrophenone 46 Pentanophenone 47 Hexanophenone 48 Heptanophenone 49 Octanophenone 50 Nonanophenone 5 1 Decanophenone 52 Undecanophenone 53 Chlorobenzene 54 1,2-Dichlorobenzene 55 1,3-Dichlorobenzene 56 1,4-Dichlorobenzene 57.84 67.32 67.32 67.32 2.808 3.532 3.532 3.532

276.337 278.353 278.353 276.337 92.140 106.167 120.194 120.194 134.221 148.247 162.274 176.301 190.328 204.355 218.381 120.151 134.177 148.204 162.231 176.258 190.285 204.311 218.388 232.365 246.392 112.559 147.004 147.004 147.004

247.198 259.435 259.277 248.683 100.456 1 17.067 134.032 133.897 150.713 167.564 184.315 201.171 217.884 234.706 25 1.540 119.403 136.200 153.052 169.817 186.582 203.4 15 220.180 236.945 253.710 270.484 97.300 110.741 110.911 110.844

272.98 1 296.969 299.886 279.958 131.549 152.079 172.293 173.691 195.232 217.077 239.063 260.952 283.590 305.2 18 326.927 153.029 174.031 196.058 2 17.968 240.137 261.523 283.668 305.856 327.255 349.826 126.293 141.224 142.517 142.627

8.032 8.030 7.49 1 8.138 9.333 9.298 9.388 9.297 9.298 9.301 9.317 9.300 9.299 9.299 9.301 9.934 9.925 9.925 9.922 9.926 9.924 9.924 9.924 9.924 9.925 9.560 9.602 9.686 9.523

6.188 6.06 1 6.197 6.197 2.778 3.189 3.614 3.613 4.052 4.496 4.933 5.357 5.764 6.1 19 6.433 1.936 2.346 2.768 3.197 3.625 4.009 4.393 4.75 1 5.035 5.344 2.826 3.295 3.324 3.393

8900 10099 9885 904 1 4510 5217 6159 6107 7002 7847 8715 9580 10458 11305 12116 5546 6453 7313 8153 9035 9915 10772 11640 12507 13371 4047 4744 4744 4745

1.656 1.884 1.950 1.944

76.80 76.80 76.80 86.28 86.28 86.28 95.76 105.24

-

57 1,2,3-Trichlorobenzene 58 1,2,4-TrichIorobenzene 59 1,3,5-Trichlorobenzene 60 1,2,3,4-Tetrachlorobenzene 6 1 1,2,3,5-Tetrachlorobenzene 62 1,2,4,5-Tetrachlorobenzene 63 Pentachlorobenzene 64 Hexachlorobenzene 65 2-Chlorotoluene 66 3-Chlorotoluene 67 4-Chlorotoluene 68 2-Bromotoluene 69 4-Iodotoluene 70 1-Chloro-4-iodobenzene 7 1 1-Bromo-3-chlorobenzene 72 2-Bromo-4-chloroacetophenone 73 Bromobenzene 74 Iodobenzene 75 Methyl benzoate 76 Isopropyl benzoate 77 Butyl benzoate 78 Isopentyl benzoate 79 p-Xylene 80 1,4-Dimethoxybenzene 8 1 Phenethyl acetate 82 Benzyl ethyl ether 83 2-Methylnaphthalene 84 2,4-Dimethylnaphthalene 85 1-Chloronaphthalene 86 Propanol 87 Butanol 124.174 124.307 124.505 137.542 137.770 137.730 150.870 164.037 114.010 114.119 114.140 121.894 129.514 126.392 118.912 154.618 105.336 112.716 128.120 161.823 78.537 95.268 17.182 34.452 44.920 25.900 44.195 160.782 140.961 70.540 87.399 156.074 157.548 158.784 170.685 172.135 172.368 185.175 197.770 146.978 148.096 148.082 153.533 162.180 156.666 149.925 194.224 133.720 140.247 164.986 206.896 229.265 248.769 153.244 173.357 187.287 163.363 178.866 198.828 171.558 104.025 125.838 9.785 9.653 9.921 9.734 9.763 9.656 9.786 9.91 1 9.415 9.427 9.297 9.473 9.410 9.621 9.714 9.914 9.604 9.645 10.020 9.977 9.997 9.991 9.059 8.869 9.493 8.961 8.617 8.503 8.781 10.847 10.845 52.40 4.197 4.157 4.197 4.944 4.944 4.944 5.691 6.438 3.40 3.40 3.40 3.61 3.93 4.06 3.74 2.57 3.017 3.334 2.15 3.09 3.74 4.15 3.10 2.23 2.36 2.64 3.70 4.20 3.80 0.27 1 0.801 3.771 3.828 3.889 4.256 4.305 4.352 4.736 5.176 3.236 3.21 1 3.206 3.606 3.918 4.110 3.693 3.232 3.217 3.590 1.969 2.776 3.228 3.561 3.224 1.804 2.377 2.190 4.068 4.429 4.105 0.362 0.922 545 1 5466 5480 6180 6208 6209 6938 7667 5079 5217 5217 522 1 5432 5005 4915 7367 4208 4292 6205 7976 8719 9603 5684 6645 7293 5910 6013 7056 5619 4222 5042

181.449 181.449 181.449 2 15.894 21 5.894 215.894 250.339 284.784 126.585 126.585 126.585 171.036 218.037 238.455 191.455 233.492 157.010 204.0 10 136.150 164.204 178.230 192.257 106.167 138.170 150.177 122.166 142.200 156.227 162.618 60.096 74.122

Table 6.4 Continued

VWV"

h ,

No. Chemical

Mr

SAh
VWVf

I' P
BlogP" 62.63 72.86 83.09 93.32 1.331 1.861 2.391 2.921
-

-AHe

logP

AH^

113.78
-

3.98 1
-

134.24
-

5.041
-

154.70
-

6.701
-

175.16
-

7.101 0.93 1.49 2.06 1.57 0.78 1.28 0.88

-

- 0.07

88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 1 10 11 1
-

Pentanol Hexanol Heptanol Octanol Nonanol Decanol Undecanol Dodecanol Tridecanol Tetradecanol Pentadecanol Hexadecanol Benzyl alcohol Cinnamanol Indole Coumarin Benzamide Salicylamide Phenacetin Uracil Caffeine Nitrobenzene 1-Fluoro-2,4-dinitrobenzene Benzyl methyl ketone 1.81 1.67 1.75

88.149 102.176 116.203 130.230 144.256 158.283 172.310 186.337 200.364 214.390 228.417 242.444 108.140 120.15 1 117.150 146.145 121.138 137.138 179.2 18 112.088 194.193 123.111 186.099 120.151 104.174 121.084 137.862 154.627 171.395 188.472 204.994 222.059 238.703 255.670 272.009 289.335 108.505 119.356 112.880 127.931 114.245 121.914 173.529 90.666 167.556 106.298 132.899 119.549

147.423 169.7 12 191.712 213.447 234.930 257.55 1 278.645 301.238 322.333 345.123 365.820 389.083 141.102 153.475 142.179 158.467 148.148 155.779 222.040 121.421 210.514 137.585 169.714 153.022 1.497 2.093 2.670 3.259 3.791 4.347 4.800 5.280 5.634 6.046 6.294 6.603 1.848 2.038 1.602 1.665 1.212 0.718 1.566 - 0.588 0.326 1.987 1.685 1.938

10.851 10.846 10.851 10.849 10.851 10.845 10.852 10.847 10.853 10.848 10.852 10.846 9.362 8.588 8.405 9.461 9.948 9.509 8.348 9.972 8.964 10.557 11.264 9.938 591 1 6786 766 1 8535 9406 10279 11157 12020 12880 13760 14636 15515 5145 5769 4705 5403 5462 6029 5403 4565 9068 4876 6945 5518

"van der Waals volume (A3 mol-'); "surface area $A2) and 'partition coefficient are calculated by MOPAC-BlogP from CACheTM;'ionization potential (eV) and 0 'enthalpy are calculated by MOPAC from CAChe M;-fvander Waals volume calculated by Bondi's method; partition coefficient calculated by Rekker's method; " enthalpy measured by liquid chromatography.

Table 6.4B Molecular properties of acidic compounds

.
rb

No. Chemical
VWV" VWVJ

w
SAb 1.980 1.683
-

IP'
BlogP" -AH' logF pK,'

-AHh

' 2
ff'

53.88 65.03 65.03 65.03 76.18 76.18 76.18 76.18 76.18 76.18 1.910 1.926 1.872 1.880 1.941 1.906 1.889
-

1.54 2.05 2.05 2.05 2.49 2.49 2.49 2.49 2.49 2.49
-

10.02 10.32 10.09 10.27 10.54 10.60 10.41 10.63 10.36 10.19 3.02 3.02 3.02 1.947 1.940 1.898
-

87.33 87.33 87.22

-

3 3 2.
10.88
-

.:
98.48
-

3.55
-

1.960

t $'

$ 2
5

112 Phenol 113 2-Methylphenol 114 3-Methylphenol 115 4-Methylphenol 116 2,3-Dimethylphenol 117 2,4-Dimethylphenol 118 2,5-Dimethylphenol 119 2,6-Dimethylphenol 120 3,4-Dimethylphenol 121 3,5-Dimethylphenol 122 2,3,4-Trime thy lphenol 123 2,3,5-Trimethylphenol 124 2,3,6-Trimethylphenol 125 2,4,6-Trirnethylphenol 126 2,3,4,5-Tetramethylphenol 127 2,3,5,6-Te tramethylphenol 128 Pen tamet hy lphenol 129 2-Ethylphenol 130 3-Ethylphenol 131 4-Ethylphenol 132 2-Chlorophenol 133 3-Chlorophenol 134 4-Chlorophenol 135 2,3-Dichlorophenol 136 2,4-Dichlorophenol 137 2,5-Dichlorophenol 138 2,6-Dichlorophenol 91.738 108.327 108.432 108.447 124.810 125.003 125.035 124.8 16 125.229 125.339 141.221 141.587 141.094 141.518 157.430 157.741 173.376 124.897 124.979 125.084 105.235 105.249 105.220 118.717 118.954 118.947 118.595 120.577 140.162 142.265 142.192 160.190 162.149 162.359 159.997 161.752 163.956 178.530 181.784 179.555 182.671 197.433 197.874 2 12.220 160.104 162.093 161.793 136.109 136.77 1 137.032 150.720 152.192 152.222 149.937 1.295 1.699 1.706 1.703 2.163 2.162 2.141 2.138 2.146 2.139 2.562 2.578 2.568 2.597 2.921 2.922 3.205 2.099 2.096 2.118 1.743 1.731 1.773 2.203 2.252 2.304 2.278 75.26 75.26 75.26 63.03 63.03 63.03 72.51 72.51 72.51 72.51 2.58 2.58 2.58 2.27 2.27 2.27 2.93 2.93 2.93 2.93 1.877 1.838 1.886 1.899 2.118 2.085 2.06 1 2.029 2.03 1 2.032 4115 5157 5238 5304 6237 6298 6312 6316 9186 6442 7158 7307 7222 7433 8220 8330 9126 5922 5984 5975 4775 4808 4810 5560 5521 5512 5548 9.1 14 8.959 9.041 8.88 1 8.924 8.78 1 8.896 8.887 8.803 8.975 8.726 8.853 8.785 8.783 8.633 8.763 8.609 8.998 9.017 8.862 9.259 9.335 9.125 9.430 9.230 9.303 9.374

94.1 13 108.140 108.140 108.140 122.166 122.166 122.166 122.166 122.166 122.166 136.193 136.193 136.930 136.193 150.220 150.220 164.247 122.166 122.166 122.166 128.558 128.558 128.558 163.003 163.003 163.003 163.003

10.2 9.9 10.0 8.48 9.02 9.38 7.45 7.89 7.50 6.79

(continued) ! 2

w
h )

Table 6.4B Continued

VWV"
Slog@

No. Chemical

Mr

SAh

IP"
-AHe
VWVf pK:

logp 2.133 2.209 2.083 2.374 2.093 2.157

-

-AHh

2.335 2.470 2.25 1 2.439

-

8.39 8.18 7.59 7.23 6.12 7.33 6.42 7.74 6.96 5.44 5.26

2.039 2.024 2.021

-

72.51 72.51 81.99 81.99 81.99 81.99 81.99 81.99 91.47 91.47 100.95 74.51 74.51 74.51 66.48 66.48 66.48 79.08 79.08 67.11 68.16 68.16 81.39 81.39 81.39 82.44
-

8.44 9.03 9.36 7.8 6.6 7.23 8.40 7.15 4.09 5.22 3.71 5.43
-

139 3,4-Dichlorophenol 140 3,5-Dichlorophenol 141 2,3,4-Trichlorophenol 142 2,3,5-TrichlorophenoI 143 2,3,6-Trichlorophenol 144 2,4,5-TrichlorophenoI 145 2,4,6-Trichlorophenol 146 3,4,5-Trichlorophenol 147 2,3,4,5-Tetrachlorophenol 148 2,3,5,6-Tetrachlorophenol 149 Pentachlorophenol 150 2-Chloro-5-methylphenol 151 4-Chloro-2-methylphenol 152 4-Chloro-3-methylphenol 153 2-Bromophenol 154 3-Bromophenol 155 4-Bromophenol 156 2,4-Dibromophenol 157 2,6-Dibromophenol 158 2-Nitrophenol 159 3-Nitrophenol 160 4-Nitrophenol 161 2,4-Dinitrophenol 162 2,5-Dinitrophenol 163 2,6-Dinitrophenol 164 3,4-Dinitrophenol 165 2-Hydroxyacetophenone 166 4-t-Butylphenol 167 Prowl 4-hvdroxvbenzoate 118.826 152.197 118.965 153.141 132.121 165.377 133.301 167.060 131.961 164.831 132.324 167.370 132.193 166.307 132.273 166.399 145.518 179.888 145.328 179.604 158.497 192.012 121.556 154.549 121.654 155.526 122.025 157.213 112.895 141.882 113.321 144.179 113.358 144.363 133.715 162.464 134.337 164.099 113.337 146.167 114.417 148.348 114.403 147.925 136.132 172.035 136.159 171.927 135.574 169.758 137.036 175.184 126.970 161.167 158.057 198.996 169.667 219.136 9.280 9.538 9.380 9.484 9.535 9.322 9.390 9.422 9.457 9.629 9.574 9.127 9.023 9.03 1 9.301 9.348 9.190 9.461 9.444 9.952 9.948 10.072 10.757 10.613 10.656 10.718 9.319 8.893 9.5 16 2.221 2.276 2.684 2.740 2.735 2.775 2.802 2.722 3.170 3.238 3.640 2.140 2.176 2.189 2.197 2.229 2.148 3.001 2.956 1.644 1.571 1.723 1.221 1.420 1.167 0.849 1.461 2.887 2.488 5514 5548 6244 6251 6293 6282 6286 6243 6974 6978 7748 5860 5891 5852 4938 4971 4954 5852 5831 5640 5655 5584 7105 7105 7172 7167 6265 7599 8719 2.93 2.93 3.68 3.68 3.68 3.68 3.68 3.68 4.43 4.43 5.17 2.71 2.71 2.71 2.48 2.48 2.48 3.35 3.35 1.92 1.27 1.27 1.58 1.58 1.58 0.93 1.21 3.60 2.67

163.003 163.003 197.448 197.448 197.448 197.448 197.448 197.448 23 1.893 23 1.893 266.338 142.585 142.585 142.585 173.009 173.009 173.009 25 1.905 25 1.905 139.110 139.110 139.110 184.108 184.108 184.108 184.108 136.150 150.220 180.203

e z

9
rn

65.36

2.378

9 2

2.
-

4.200 4.320 4.450 5 3.940 .2 3.250 g 4.580 4.070 % 2.821 4.470 4.090 E 4.480 8'

2:
4.270 4.340
-

?t

3.20 3.24 1.oo 1.oo 1.oo 2.86 2.86 2.45 1.94 0.72 0.99 0.94 1.30 1.28 1.37 2.18 1.31 1.99 1.86 2.297 2.297 2.297 2.571 2.465 2.428
-

T
4.350
-

2
-

4
4.440 4.340
(continued)

s
2.539 2.623 2.623 2.273 2.486 2.513 2.393

168 Butyl4-hydroxybenzoate 169 4-Chloro-3,5-dimethylphenol 170 1,2-Dihydroxybenzene 171 1,3-Dihydroxybenzene 172 1,4-Dihydroxybenzene 173 1-Hydroxynaphthalene 174 2-Hydroxynaphthalene 175 1-Hydroxy-2,4-dinitronaphthalene 176 Benzoic acid 177 3,4,5-Trihydroxybenzoic acid 178 3,4-Dihydroxybenzoic acid 179 3,5-Dihydroxybenzoic acid 180 2,4-Dihydroxybenzoic acid 181 4-Hydroxybenzoic acid 182 3-Hydroxybenzoic acid 183 2-Hydroxybenzoic acid 184 4-Hydroxy-3-methoxybenzoic acid 185 3-Methoxybenzoic acid 186 4-Methoxybenzoic acid 187 2-Methylbenzoic acid 188 3-Methylbenzoic acid 189 4-Methylbenzoic acid 190 2-Ethylbenzoic acid 191 3-Ethylbenzoic acid 192 4-Ethylbenzoic acid 193 2,4-Dimethylbenzoic acid 194 2,5-Dimethylbenzoic acid 195 2,6-Dimethylbenzoic acid 196 3,4-Dimethylbenzoic acid 197 3,5-Dimethylbenzoic acid 198 2,4,6-Trimethylbenzoic acid 186.310 240.792 138.449 177.464 99.793 130.523 99.888 131.163 99.852 131.027 135.452 166.326 135.586 167.716 178.767 212.8 16 110.724 142.853 134.292 171.289 126.668 163.122 126.699 164.064 126.388 161.686 118.751 153.441 118.740 153.334 118.396 151.204 143.959 184.413 135.989 174.473 136.002 174.655 126.868 160.859 127.432 164.572 127.336 164.514 143.994 182.946 143.930 184.475 143.963 184.244 143.412 181.614 143.428 181.896 143.334 179.269 144.135 186.656 144.135 186.056 159.835 200.682 9.520 8.892 8.796 9.044 8.736 8.465 8.643 9.792 10.083 9.372 9.230 9.407 9.505 9.605 9.519 9.506 9.100 9.391 9.479 9.728 9.747 9.824 9.78 1 9.707 9.784 9.636 9.446 9.520 9.636 9.581 9.488 2.930 2.624 0.878 0.853 0.878 2.618 2.612 2.173 1.612 0.756 0.992 0.968 0.765 1.260 1.361 1.146 1.501 1.825 1.688 2.038 2.020 2.014 2.390 2.422 2.413 2.423 2.448 2.429 2.472 2.472 2.808 76.51 76.51 76.51 86.74 86.74 86.74 2.827 87.66 87.66 87.66 87.66 87.66 98.81 958 1 689 1 5004 4836 4915 5637 5613 8717 5159 7595 6773 6681 6652 5874 5916 5967 7643 6799 6790 6095 6343 6325 71 12 7042 7042 7255 7254 71 15 7294 7518 8298

194.230 156.612 110.112 110.112 110.112 144.173 144.173 234.168 122.123 170.121 154.122 154.122 154.122 138.123 138.123 138.123 168.149 152.149 152.149 136.150 136.150 136.150 150.177 150.177 150.177 150.177 150.177 150.177 150.177 150.177 164.204

." 1

c .
h )

Table 6.4B Continued

VWV"
VWVf
-

No. Chemical
-

M r

SAb

IF
BlogP'

-AHe
logP

-AHh

pKai

2.5 17
-

3.380 3.960 3.960

-

3.600 3.460 2.726 1.94

-

3.810 3.980 4.300

-

74.84 74.84 74.84 84.32 84.32 84.32 84.32 84.32 77.96 77.96 77.96 79.55 86.74 86.74 86.74
-

2.566 2.704 2.718 2.862 2.918 2.619 3.008 2.964 2.677 2.797 2.828 2.428 2.441 2.486
-

4.325
-

85.07 85.07 85.07

-

2.599
-

2.777
-

n
-

199 200 201 202 203 204 205 206 207 208 209 2 10 21 1 2 12 2 13 2 14 2 15 2 16 217 218 219 220 221 222 223 224 225 226 227 159.776 159.033 159.022 175.267 176.071 174.850 173.886 175.409 166.909 166.492 166.737 163.937 182.111 184.65 185.405 201.187 204.385 204.073 177.038 178.739 178.664 181.183 182.608 171.653 173.179 203.920 194.122 193.786 195.496 1.991 2.135 2.143 2.538 2.495 2.398 2.655 2.607 2.300 2.491 2.488 2.037 2.413 2.443 2.470 2.776 2.846 2.838 2.564 2.568 2.495 1.498 1.360 1.726 1.807 1.872 2.111 2.253 1.145
-

2-Chlorobenzoic acid 156.568 3-Chlorobenzoic acid 156.568 4-Chlorobenzoic acid 156.568 2,4-Dichlorobenzoic acid 191.013 2,5-Dichlorobenzoic acid 191.013 2,6-Dichlorobenzoic acid 191.013 3,4-Dichlorobenzoic acid 191.013 3,5-Dichlorobenzoic acid 191.013 2-Bromobenzoic acid 201.019 3-Bromobenzoic acid 201.019 4-Bromobenzoic acid 136.150 Phenylacetic acid 136.150 2-Methylphenylacetic acid 150.177 3-Methylphenylacetic acid 150.177 4-Methylphenylacetic acid 150.177 2-Ethylphenylacetic acid 164.204 3-Ethylphenylacetic acid 164.204 4-Ethylphenylacetic acid 164.204 2-Chlorophenylacetic acid 170.595 170.595 3-Chlorophenylacetic acid 170.595 4-Chlorophenylacetic acid 168.149 3,4-Dihydroxyphenylaceticacid 168.149 2,5-Dihydroxyphenylacetic acid 152.149 2-Hydroxyphenylacetic acid 152.149 4-Hydroxyphenylacetic acid 4-Hydroxy-3-methoxyphenylacetic acid 182.176 166.176 3-Methoxyphenylacetic acid 166.176 4-Methoxyphenylacetic acid 184.148 3,4-Dihydroxymandelic acid 9.93 1 9.946 10.023 10.066 9.847 9.97 1 9.989 10.010 9.952 9.969 9.824 9.682 9.476 9.443 9.383 9.436 9.361 9.310 9.643 9.622 9.518 8.890 8.810 9.195 9.106 8.752 9.1 11 8.992 8.937 5930 5870 5864 6647 6666 6674 6591 6609 6122 6045 6047 6010 7073 7188 7184 7776 8390 7885 6751 7124 7124 7667 7620 6783 6788 8539 7675 7709 8638 2.473 0.98 0.80 1.47 1.26 1.18 1.99 1.95 0.144

124.325 124.254 124.250 137.827 137.947 137.710 137.721 137.913 132.145 132.23 132.209 127.527 143.730 144.013 144.143 160.255 160.509 160.492 140.569 140.759 140.631 143.030 143.039 134.959 135.192 160.492 152.512 152.489 151.882

4,216 4.387 3.964 4.225 4.433 4.394 4.261 4.394 3.493

80.18 1.650
-

3
$
%

82.32 2.710
-

0.84 1.33 1.15 1.46 2.32 2.37

-

1.51
-

3.380 3.507 3.330 4.545 4.630 4.380 4.307

s. s
2.789 2.601 2.609 4.558 4.492 4.673 4.579 4.719

93.47

9 2
5
-

86.04 96.05
-

2.063 2.250 2.522

-

9
1.92
-

4.590
-

e
-

2.
3 :

91.65 114.11 121.34 96.15

-

3.033 3.28 1 3.388 1.518 1.55

-

3.506 S
-

b
-

228 229 230 23 1 232 233 234 235 236 237 238 239 240 24 1 242 243 244 245 246 247 248 249 135.783 168.774 160.809 154.457 146.507 138.435 163.810 155.116 171.801 155.207 189.048 144.183 160.924 177.872 139.965 156.451 175.527 158.769 166.550 123.569 154.456 162.354 173.022 2 13.274 203.41 1 197.932 188.227 177.359 209.030 198.854 219.02 1 199.577 239.78 1 185.755 207.898 229.559 176.493 195.771 219.961 203.669 212.370 159.147 190.262 199.135 9.991 8.784 9.088 8.947 9.1 12 9.472 9.2 16 9.259 8.854 9.253 8.755 9.520 9.434 9.41 1 8.825 8.445 8.431 10.020 9.569 9.165 9.063 8.903 1.752 6650 1.551 9369 1.892 8558 1.762 8310 2.016 7416 2.333 6676 2.539 8326 2.751 7826 2.230 9159 2.765 7851 2.616 10039 2.454 6948 2.890 7853 3.327 8727 0.864 6419 1.380 7416 1.604 8407 1.573 8184 1.038 8802 - 1.064 6557 2.820 6766 2.311 7575

Mandelic acid 4-Hydroxy-3-methoxymandelic acid 3-Methoxymandelic acid 3,4-Dihydroxycinnamic acid 4-Hydroxycinnamic acid trans-Cinnamic acid 3-Methoxycinnamic acid 4-Methyl( trans)cinnamic acid 4-Hydroxy-3-methoxycinnamic acid 4-Methylcinnamic acid 4-Methoxy-3-methoxycinnamic acid Phenylpropionic acid Phenylbutyric acid Pheny lpen tanoic acid 3-Indoleacetic acid 3-Indolepropionic acid 3-Indolebutyric acid Hippuric acid 2-Hydroxyhippuric acid Uric acid 2-Naphthoic acid 3-Hydroxy-2-naphthoic acid

152.149 198.175 182.176 180.160 164.160 148.161 178.187 162.188 194.187 162.188 208.213 150.177 164.204 178.230 161.160 175.187 189.213 179.175 195.174 168.112 172.183 188.182

; i :

van der Waals volume mol-); surface area $A2) and partition coefficient are calculated by MOPAC-BlogP from CACheTM;ionization potential (eV) and enthalpy are calculated by MOPAC from CAChe M;.fvander Waals volume calculated by Bondis method; partition coefficient calculated by Rekkers method; enthalpy measured by liquid chromatography; dissociation constant.

(A

2-

126

Chapter 6

The ionization potential calculated by MOPAC seemed to be capable of correcting the log k values of polychlorobenzenes when alkylbenzenes were used as standards, as shown in Figure 6.5 where the ionization potentials of alkyl and polychlorobenzenes are shown. The relationship between the van der Waals volume ( V ) and the logk values of the alkylbenzenes is given by: logk=Ax V + B (6.9)

The constants A and B are related to the percentage concentration of acetonitrile (x), and their values are:
A = 1.1503 x

+ 4.9525 x 10-5x - 1 .Of325 x 10-6x2 B = 1.305 - 3.2936 x lO-'x + 1.2900 x 10-4x2

(r2 = 1 .OOO)

(r2 = 1.000)

For hypothetical alkylbenzenes having the same van der Waals volume as alkanes, polycyclic aromatic hydrocarbons, and polychlorobenzenes, the calculated logk values were determined by means of Equation 6.9 with constants A and B obtained from the alkylbenzene results. The difference (A logk) between the measured and the calculated values of logk in the same eluent was determined. The ionization potential energy, IP, of alkylbenzenes calculated by MOPAC was used as the standard, and it was almost constant for a series of alkyl compounds. The difference in the ionization potential, AIP, of other compounds from that of alkylbenzenes was calculated. The relationship between A log k and AIP is given by: Alogk
=

CAIP

+D
1

(6.10)

10.0 1

J."'

'

50

100

150

200

250

van der Waals volume

A3/ mole

Figure 6.5

Relationship of ionization potential to van der Waals volume. Compounds: W , alkylbenzenes; and , polychlorobenzenes. Number beside symbols: see Table 6.4.

Influence o Physical Chemistry on Separations in Liquid Chromatography f

127

where C and D are constants in the same eluent. The relationship between C, D, and the acetonitrile concentration (x)was obtained from

C = 0.266

+ 2.826 x 10-3x - 3;640 x 10-5x2

(? = 0.999)
(r2 = 0.989)

(6.1 1) (6.12)

D = -0.1 13 + 2.177 x 10A3x 8.798 x 10A6x2 -

The log k of alkanes, polycyclic aromatic hydrocarbons, and polychlorobenzenes was therefore estimated from Equations 6.9 and 6.10 in 60-90% aqueous acetonitrile. The calculated equation is: log kcalc = A X

+ B + CAIP + D

(6.13)

where A , B, C, and D are constants. The correlation coefficient between the measured and calculated retention factors was -0.922 (n = 27), 0.946 (n = 32), 0.966 (n = 3 9 , and 0.971 (n = 37) in 60, 70, 80, and 90% aqueous acetonitrile, respectively. This method was applied to the prediction of the retention factor of phenols. First, this method requires a relationship between logk and the van der Waals volume of a homologous series of alkyl compounds such as alkylbenzenes or alkylphenones. When alkylbenzenes are used as the standard, the value of B in Equation 6.8 should be altered to give a parallel relationship at a suitable position ( E )on the y-axis for a different group of compounds. When B is moved to E for phenols, the logk values of para-alkylated phenols 01) are simply predicted from their van der Waals volume by the following equation: logkcalc = Ax

+E

(6.14)

For other substituted phenols, the difference in their ionization potential from that of phenol was calculated, and then their Alogk was obtained from Equation 6.10. The logk of substituted phenols was then estimated from Equations 6.9 and 6.13. The retention time of phenols was predicted in 70 and 60% acidic aqueous acetonitrile on an ODS silica gel column. The constants A , B, C, and D were obtained from the above equations. The result in 60% aqueous acetonitrile is shown in Figure 6.6. The correlation coefficients between the measured and predicted retention factors of substituted phenols in 60 and 70% acidic aqueous acetonitrile were 0.974 (n = 36) and 0.967 (n = 36), respectively. In this system, the values of the slopes, which indicate the relationship between the measured and predicted retention factors, were 0.81 and 0.94 in 60 and 70% acidic aqueous acetonitrile, respective~y.~~ This system of using the van der Waals volume and ionization potential is much simpler than previous systems. It requires only one homologous series of alkyl compounds and one standard such as phenol for substituted phenols. The precision could be improved for polar compounds if para-alkylphenols and para-alkylbenzoic acids are used; however, only a few such compounds are

128

Chapter 6

149

0
148

0
Old7

112

O!

measured retention factor Figure 6.6 Comparison of measured and predicted retention factors of substitute phenols. Column, 5 pm ODs-bonded silica gel, 15 cm x 4.6 m m i.d.; eluent, 60% aqueous acetonitrile containing 0.05 M phosphoric acid. Number beside symbols: see Table 6.4.

available at present. This system seems to be practical for estimating the approximate retention time.

3 Enthalpy as a Thermodynamic Property in Retention Studies

The chromatographic retention parameters used in correlation studies are normally assumed to be proportional to the free energy change. Adsorption is accompanied by a loss of surface free energy. Free energy change is related to enthalpy, AH, and entropy, AS, changes by the Gibbs equation: AG
=

A H - TAS

(6.15)

The retention factor is related to the thermodynamic equilibrium constant K for solute binding by k = PK, where P is the phase ratio of the column. The freeenergy change for the chromatographic process is expressed by AG
=

-RTlnK

-RTln(k/P)

(6.16)

where R is a gas constant (8.314 J K-' mol-') (1 cal is 4.184 J). From Equations 5.38 and 5.39: Ink
=

-(dH/RT)

+ ( d S / R ) + In /?

(6.17)

A van't Hoff plot of Ink against 1/T yields a straight line if the stationary phase is not deformed. The relation Ink and - A H / R T is used to study the tempera ture-dependen t selectivity of the stationary phase.33-3

Influence o Physical Chemistry on Separations in Liquid Chromatography f

129

Measurement of Enthalpy by Liquid Chromatography

Enthalpy can be measured by liquid chromatography where enthalpy is a slope of the relationship between Ink and the inverse value of the absolute temperature. A schematic diagram is shown in Figure 6.7. The slope depends upon the solutes being retained by the same liquid chromatographic mechanism. An example is given in Table 6.4. The results, measured on an octadecyl-bonded vinyl alcohol copolymer gel, did not show a simple linear relationship. This is due to a conformation change of the octadecyl-bonded vinyl alcohol copolymer gel stationary phase material, which has a phase transition point at about 33 "C.

Enthalpy Related to Retention

The higher the enthalpy change ( A H ) , the longer the retention, and log k shows a linear correlation to e n t h a l ~ yAs seen in Figure 6.8, a number of groups of .~~ compounds exhibit such a linear relationship between enthalpy and log k, although it is not perfect. The AH values were independent of their van der Waals volumes, as seen in Figure 6.9. The AAH value of a methylene unit of alkanols is fairly constant.36 However, the AAH value of a methylene unit of alkylbenzenes was not constant and the larger molecules had a higher AAH value than the average AAH value of a methylene unit. This larger value may be due to dimerization between solutes as described ~chematically,~~ to a strong direct interaction between alkyland/or benzenes and the octadecyl group of the stationary phase material, because alkylbenzenes with a longer alkyl group must lose their aromaticity. The isomeric effect of the AAH value of substituted chlorobenzenes was very small. This means that the molecular size of the isomers is almost the same; however, the actual size is not the same under these chromatographic condition~.~~

AH = slope = b / a
1/ T

r e c i p r o c a l of a b s o l u t e t e m p e r a t u r e
Figure 6.7 Measurement of enthalpy using chromatography for the relationship between absolute temperature and retention factor.

130
4000
2 5

Chapter 6

Q)

9 5 B

ox4

40

3000
64

I p

I p

> I P

2000

5 C
a
1O W AA
87

0
10
I

32
I

0.5

1.5

log k
Figure 6.8 Relationship of enthalpy to log k values. Column, O D s silica, ERC-ODs, 15 cm x 6.0 mm i.d.; eluent, 80% aqueous acetonitrile at 30C. Numbers beside symbols: see Table 6.4. 0 ,Polycyclic aromatic hydrocarbons; x ,alkylbenzenes; 0, halogenated benzenes; A, alkanols; and 0, alkanes.

4000
Q)

,
40 @ 2 5

2 i :
Q

4 o

<Ag5

3000-

>r a - **5
(I I

Q)

100

A
75 125 1 5

100025

87AA

van der Waals volume cm3/mole

Figure 6.9 Relationship of enthalpy to Bondi's van der Waals volumes. Column, ODs silica, ERC-ODs, 15 cm x 4.6 mm i.d.; eluent, 80% aqueous acetonitrile. Numbers beside symbols: see Table 6.4. 0 ,Polycyclic aromatic hydrocarbons; x ,alkylbenzenes; 0, halogenated benzenes; A, alkanols; and 1, alkanes.

The enthalpy of methylphenols was about 2.0 kcal mol-', and that of chlorophenols varied from 2.0 to 2.4 kcal mol- in the case of pentachlorophenol, indicating that the retention difference depended not upon the size but on the n-electron density.39 A similar result was obtained for alkylated and halogenated aromatic acids, whose enthalpies were nearly equal, but whose retention factors were different.40 The AH values may depend on the type of stationary phase used and the water content of the eluent.41

'

Influence of Physical Chemistry on Separations in Liquid Chromatography

131

The highest enthalpy was found on non-endcapped high carbon octadecylbonded silica gels, and was slightly higher on high-carbon endcapped gels than on low-carbon endcapped gels. There was a selectivity difference in the retention between alkanols and other hydrophobic compounds that depended on the endcapping treatment applied.42 The order of enthalpies of phenols showed no direct relationship to the retention factor or hydrophobicity of octadecylbonded silica gels with different carbon loadings.43 Further, enthalpy was not directly related to the log k values nor to the alkyl chain length of bonded-silica gels, although the retention behaviour of individual groups of compounds such as polycyclic aromatic hydrocarbons, and especially the retention of smallersize polar compounds, appeared to indicate this.44

4 Conclusion
The qualitative analysis of retention behaviour in liquid chromatography has now become possible. Quantitative retention-prediction is, however, still difficult; the prediction of retention time and the optimization of separation conditions based on physicochemical properties have not yet been completely successful. One reason is the lack of an ideal stationary phase material. The stationary phase material has to be stable as part of an instrument, and this is very difficult to achieve in normal-phase liquid chromatography because the moisture in organic solvents ages the silica gel. Several improved stationary phase materials have been synthesized for reversed-phase liquid chromatography. One material is vinyl alcohol copolymer gel. This stationary phase is quite polar and chemically very stable; however, it demonstrated a strong retention capacity for polycyclic aromatic hydrocarb o n ~ . Although stable octadecyl- and octyl-bonded silica gels have been ~ ~ ~ synthesized from pure silica and are now commercially available, such an optimization system has not yet been built. Further experiments are required to elucidate the retention mechanism, and to systematize it within the context of instrumentation.

5 References
1. A. Leo, C. Hansch, and D. Elkins, Chern. Rev., 1971,71, 525. 2. R.F. Rekker, The Hydrophobic Fragmental Constant, Elsevier Scientific, Amsterdam, 1977. 3. W.J. Haggerty and E.A. Murrill, Res. Dev., 1974, 25, 30. 4. R.M. Calson, J. Chromatogr., 1975, 107, 219. 5. J.M. McCall, J. Med. Chem., 1975, 18, 549. 6. M.S. Mirrless, S.T. Moulton, C.T. Murphy, and P.J. Taylor, J. Med. Chem., 1976, 19, 615. 7. K. Miyake and H. Terada, J. Chromatogr., 1978, 157, 386. 8. M.J. OHare and E.C. Nice, J. Chromatogr., 1979, 171, 209. 9. J. Grunenberg and R. Herges, J. Chem. If. Comput. Sci., 1995,35,905. 10. C. Yamagami, M. Yokora, and N. Takao, J. Chrornatogr., 1994,662,49. 11. A. Kakoulidou and R.F. Rekker, J . Chromatogr., 1984,295, 341. 12. M.J.M. Wells, C.R. Clark, and R.M. Patterson, J. Chromatogr. Sci., 1981, 19, 573.

132

Chapter 6

13. 14. 15. 16. 17.

18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30.

31. 32. 33. 34. 3 5.

36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51.

M.D. Amboise and T. Hanai, J. Liq. Chromatogr., 1982,5,229. K. Calmmer, L.-E. Edholm, and B.E.F. Smith, J. Chromatogr., 1977,136,45. J.F. Schabron, R.J. Hurtubise, and H.F. Silver, Anal. Chem., 1980,50, 19 1 1 . N. Nomura, Toyama Univ. Kyoikugakubu-Kiyo, 1979,27, 1. B.L. Karger, J.R. Gant, A. Harkopt, and P.H. Weiner, J. Chromatogr., 1976, 128, 65. T. Hanai, Chromatographia, 1979, 12,77. T. Hanai, K.C. Tran, and J. Hubert, HRC & CC, 1981,4,454. T. Hanai, S. Hara, S. Mori, and T. Hanai (eds), Chromatography, the Separation System, Maruzen, Tokyo, 1981, p. 138 (in Japanese). T. Hanai and J. Hubert, HRC & CC, 1983,6,20. T. Hanai and J. Hubert, J. Liq. Chromatogr., 1985,8,2463. T. Hanai and J. Hubert, J. Chromatogr., 1982,239, 527. T. Hanai and J. Hubert, HRC & CC, 1981,4,500. T. Hanai, K.C. Tran, and J. Hubert, J . Chromatogr., 1982,239,385. T. Hanai and J. Hubert, HRC & CC, 1984,7,524. D.D. Perrin, B. Dempsey, and E.P. Serjeant, pK, Prediction for Organic Acids and Bases, Chapman & Hall, London, 198 1 . A. Bondi, J. Phys. Chem., 1964,68,441. T . Hanai, J. Chromatogr., 1985,332, 189. T. Hanai, H. Hatano, N. Nimura, and T. Kinoshita, J. Liq. Chromatogr., 1993, 16, 1453. J. Yamaguchi, T. Hanai, and H. Cai, J. Chromatogr., 1988,441, 183. T. Hanai, H. Hatano, N. Nimura, and T. Kinoshita, Analyst, 1994, 119, 1167. W. Melander, D.E. Campbell, and C. Horvath, J. Chromatogr., 1978,158,215. T. Paryjczak, Gas Chromatography in Adsorption and Catalysis, Ellis Horwood Limited, New York, 1986, p. 231. R. Kaliszan, Quantitative Structure-C hromatograp hic Reten tion Rela tionships, Wiley, New York, 1987. T. Hanai and J. Hubert, J. Chromatogr., 1984, 291, 81. A. Ben-Neim, Hydrophobic Interaction, Plenum Press, New York, 1980. T. Hanai, A. Jukurogi, and J. Hubert, Chromatographia, 1984,19,266. Y. Arai, M. Hirukawa, and T. Hanai, J . Chromatogr., 1987,384,279. Y. Arai, J. Yamaguchi, and T. Hanai, J. Chromatogr., 1987,400,21. H. Colin, J.C. Diez-Mesa, G. Guiochon, T. Czaikowska, and I. Miedziak, J . Chromatogr., 1978, 167,41. J. Yamaguchi, T. Hanai, and H. Cai, J. Chromatogr., 1988,441, 183. J. Yamaguchi and T. Hanai, J. Chromatogr. Sci., 1989,27, 710. J. Yamaguchi and T. Hanai, Chromatographia, 1989,27, 371. T. Hanai, Y. Arai, M. Hirukawa, K. Noguchi, and Y. Yanagihara, J. Chromatogr., 1985,349,323. Y. Arai, M. Hirukawa, and T. Hanai, Nihon Kagaku Kaishi, 1986,969. Y. Arai, M. Hirukawa, and T. Hanai, J. Liq. Chromatogr., 1987, 10,635. M. Hirukawa, Y. Arai, and T. Hanai, J. Chromatogr., 1987,395,481. Y. Arai, M. Hirukawa, and T. Hanai, J. Chromatogr., 1987,400,27. T. Hanai, M. Ohhira, and T. Tamura, L C & GC, 1988,6,922. M. Ohhira, F. Ohmura, and T. Hanai, J . Liq. Chromatogr., 1989, 12, 1065.

Subject Index
Affinity liquid chromatography, 7,8 Amino acid, 67,74 Asymmetry (As), 39 Blood serum, 51, 52, 53 Bondi, 116 Carboxylic acid, 68 Catecholamine, 69 Cell volume (flow cell), 25 Charge-transfer, 8 Chiral separation, 7, 8 Chiration, 8 Chlorosilane, 35 Chromatographic system, 11 Column efficiency, 96, 100 Column length (L),3, 100 Column packing, 37 Column permeability (KO), 107 Column switching, 17 Column temperature, 18,80 Column test, 40 Connecting tubing (volume), 26 Counter ion, 73, 79 Dead volume ( VM), 4 Degassing, 17 Detector (detection), 19, 80 Detector sensitivity, 23 Dissociation constant (K), 113 Eddy diffusion, 102 Effective theoretical plate number (N,K), 4, 100 Eluent, 89 Enthalpy, 117, 128 Entropy, 128 Gradient, 13 Haemoglobin (hemoglobin), 70 Hamiltonian, 38 Hammet, 113 Heavy metal, 41 Height equivalent to a theoretical plate ( H ) , 3,100 Hold-up volume ( VM), 4 Hydantoin, 5 1 Injector, 11 Ion-exchange, 7,8, 36 Ion-pair liquid chromatography, 7, 70, 76 Ionic strength, 80 Ionization potential, 117, 126 Ligand-exchange, 8 Longitudinal diffusion, 103 log P,58, 110, 117 Mobile phase, 103 Molecular mass (Mr),8, 117 Molecular property, 117 Normal-phase liquid chromatography, 7, 8, 81 Organic modifier, 57,60, 80 Organic polymer, 49 Organic solvent, 57, 59,6 1, 87 Particle size (dp),4, 102 pH, 65,75 pK,, 66,79,113,121 Properties of solvent, 59 Protein, 50, 54, 55, 68,69 Quantitative structure retention relationship, 109 Rekker, 111 Recycle, 15 Resolution (Rs), 1,96, 101 Retention volume (V,), 4 Reversed-phase ion-pair, 8 Reversed-phase, 7, 8, 57 133

134 Salting-out, 8 Separation factor (a), 1,6, 100 Sieving, 37 Silica gel, 32,46 Size-exclusion liquid chromatography, 7, 8,90 Solubility, 7, 89 Solvent, 57, 59,61,85, 87 Stationary phase, 32,46, 103

Subject Index

Theoretical plate number (N), 1,4,39,97, 100 Time constant, 23 Trouble shooting and maintenance, 27, 82 Van Deemter, 106 Van der Waals, 115, 117 Vant Hoff plot, 128 Viscosity, 15 Void volume, 43