R.P.W.Scott Liquid chromatography

This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee

Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.

1

BOOK 3

Chrom-Ed Book Series

Raymond P. W. Scott

LIQUID

CHROMATOGRAPHY



2

COPYRIGHT @2003 by LIBRARY4SCIENCE, LLC

ALL RIGHTS RESERVED

Neither this book or any part may be reduced or transmitted in any form

or by any means, electronic or mechanical , including photocopying,

microfilming, and recording or by any information storage and retrieved

system without permission in writing from the publisher except as

permitted by the in-user license agreement.

World Wide Web

http://www.library4science.com/



3

Chrom-Ed Book Series

Book 1 Principles and Practice of Chromatography

Book 2 Gas Chromatography

Book 3 Liquid Chromatography

Book 4 Gas Chromatography Detectors

Book 5 Liquid Chromatography Detectors

Book 6 The Plate Theory and Extensions for

Chromatography Columns

Book 7 The Thermodynamics of Chromatography

Book 8 The Mechanism of Retention

Book 9 Dispersion in Chromatography Columns

Book 10 Extra Column Dispersion

Book 11 Capillary Chromatography

Book 12 Preparative Chromatography

Book 13 GC Tandem Systems

Book 14 LC Tandem Systems

Book 15 GC Quantitative Analysis

Book 16 Ion Chromatography

Book 17 Silica Gel and Its Uses in Chromatography

Book 18 Thin Layer Chromatography

Book 19 Chiral Chromatography

Book 20 Bonded Phases

Book 21 Chromatography Applications

COPYRIGHT @2003 by LIBRARYFORSCIENCE, LLC

ALL RIGHTS RESERVED



4

Introduction ........................................................................................... 1

The Basic Liquid Chromatograph .......................................................... 2

The Gradient Programmer and the LC Pump ..................................... 4

The Gradient Programmer ............................................................. 4

The LC Pump .................................................................................... 6

The Pneumatic Pump ..................................................................... 6

Non-Return Valves ........................................................................ 8

The Syringe Pump ......................................................................... 9

The Rapid Refill Pump ................................................................ 12

The Twin-Headed Pump. ............................................................. 13

The Diaphragm Pump.................................................................. 14

The Sample Valve ........................................................................... 16

Column Switching ....................................................................... 19

Column Ovens .................................................................................... 20

Detectors ............................................................................................. 21

The UV Detector ............................................................................. 22

The Fixed Wavelength Detector ....................................................... 23

The Multi-Wavelength Detectors ..................................................... 26

The Multi-Wavelength Dispersive Detector ................................. 26

The Diode Array Detector ............................................................... 28

The Electrical Conductivity Detector ................................................... 29

The Fluorescence Detector................................................................... 32

The Refractive Index Detector ............................................................. 34

The Tridet Multi Functional Detector .................................................. 36

Liquid Chromatography Stationary Phases .......................................... 41

Silica Gel ......................................................................................... 42

The Preparation of Spherical Silica Gel ....................................... 43

The Structure of Silica Gel ........................................................... 44

The Thermogravimetric Analysis of Silica Gel ............................. 47

Bonded Phases ................................................................................ 49

The Synthesis of Bonded Phases .................................................. 49

Bonded Phase Synthesis by Reaction in a Solvent........................ 50

The Fluidized Bed Method for Bonded Phase Synthesis .............. 53

Choosing a Bonded Phase ............................................................... 56

Types of Bonded Phase ................................................................... 57

Oligomeric Phases ....................................................................... 58

Bulk Phases ................................................................................. 59



5

Interactions Between 'Brush' and 'Bulk' Reverse Phases and

Aqueous Solvents ........................................................................ 60

The Retention Properties of Bulk and Brush Phases. ........................... 63

Macroporous Polymers ........................................................................ 66

LC Mobile Phases ............................................................................... 68

Solvent/Solute Interactions with the Silica Gel Surface .................... 69

Solute Stationary Phase Interactions ............................................ 71

Solvent/Solute Interactions with the Reversed Phase Surface ........... 74

Molecular Interactions in the Mobile Phase ..................................... 76

Aqueous Solvent Mixtures ............................................................... 78

Chiral Stationary Phases ...................................................................... 81

Macrocyclic Glycopeptide Phases .................................................... 83

Cyclodextrin .................................................................................... 89

Liquid Chromatography Applications .................................................. 93

References ......................................................................................... 101



1

Introduction

Liquid chromatography (LC) was the first type of chromatography to be

discovered and, in the form of liquid-solid chromatography (LSC), was

originally used in the late 1890s by the Russian botanist, Tswett (1), to

separate and isolate various plant pigments. The colored bands he

produced on the adsorbent bed evoked the term chromatography (color

writing) for this type of separation. Initially the work of Tswett was not

generally accepted, partly due to the original paper being in Russian and

thus, at that time, was not readily available to the majority of western

chemists and partly due to the condemnation of the method by

Willstatter and Stoll (2) in 1913. Willstatter and Stoll repeated Tswett's

experiments without heeding his warning not to use too "aggressive "

adsorbents as these would cause the chlorophylls to decompose. As a

consequence, the experiments of Willstatter et al. failed and their

published results, rejecting the work of Tswett, impeded the recognition

of chromatography as a useful separation technique for nearly 20 years.

In the late 1930s and early 1940s Martin and Synge introduced a form

of liquid-liquid chromatography by supporting the stationary phase, in

this case water, on silica gel in the form of a packed bed and used it to

separate some acetyl amino acids. They published their work in 1941 (3)

and in their paper recommended the replacement of the liquid mobile

phase with a suitable gas, which would accelerate the transfer between

the two phases and provide more efficient separations. Thus, the concept

of gas chromatography was born. In the same paper in 1941, Martin

and Synge suggested the use of small particles and high pressures in LC

to improve the separation that proved to the critical factors that initiated

the development of high performance liquid chromatography.

To quote Martin's original paper,

"Thus, the smallest H.E.T.P. (the highest efficiency) should be

obtainable by using very small particles and a high pressure difference

across the column".



2

The statement made by Martin in 1941 contains all the necessary

conditions to realize both the high efficiencies and the high resolution

achieved by modern LC columns. Despite his recommendations,

however, it took nearly fifty years to bring his concepts to fruition.

Activity in the field of liquid chromatography was eclipsed in the 1950s

by the introduction of gas chromatography and serious attempts were

not made to improve LC techniques until the development of GC neared

completion in the mid 1960s. The major impediment to the development

of LC was the lack of a high sensitive detector and it was not until the

refractive index detector was developed by A. Tiselius and D. Claesson

(4) in 1942 could the technique be effectively developed.

Tswett's original LC consisted of a vertical glass tube, a few centimeters

in diameter and about 30 cm high, packed with the adsorbent (calcium

carbonate). The plant extract pigments was placed on the top of the

packing and the mobile phase carefully added to fill the tube. The

solvent percolated through the packing under gravity, developing the

separation, which could be seen as different colored bands at the wall of

the tube. The simple apparatus of Tswett contained all the essentials to

achieve a chromatographic separation.

The contemporary chromatograph, however, is a very complex

instrument operating at pressures up to 10,000 p.s.i providing flow rates

ranging from a few microliters per minute to 10 or 20 ml/minute

depending on the type of LC that is carried out. Modern detectors can

detect solutes at concentration levels of 1x10-9 g/ml and an analysis can

be completed in a few minutes with just a few micrograms of sample.

The Basic Liquid Chromatograph

The basic liquid chromatograph consists of six basic units. The mobile

phase supply system, the pump and programmer, the sample valve, the

column, the detector and finally a means of presenting and processing

the results. A block diagram of the basic liquid chromatograph is shown

in figure 1



3

Mobile Phase Supply System

Column and Thermostat

Detector

Data Acquis ition and Processing

Pump and Programmer

Sample Valve

RecorderPrinter

Figure 1. The Basic Liquid Chromatograph

The Mobile Phase Supply System

The mobile phase supply system consists of number of reservoirs (200

ml to 1,000 ml in capacity). At least two reservoirs would be necessary

and are usually constructed of glass or stainless steel and contain an exit

port open to air. Stainless steel, however, is not considered satisfactory

for mobile phases buffered to a low pH and containing certain materials

that can cause corrosion. Each reservoir is usually fitted with a gas

diffuser through which helium can be bubbled. Many solvents and

solvent mixtures (particularly aqueous mixtures) contain significant

amounts of dissolved nitrogen and oxygen from the air. These gasses



4

can form bubbles in the chromatographic system that cause both serious

detector noise and loss of column efficiency. As helium is very insoluble

in most solvents, it purges the oxygen and nitrogen from the solvent but

does not produce bubbles in the system itself. Applying a vacuum to the

reservoir is not a permanent solution to dissolved air as, on releasing the

vacuum to allow the solvent to pass to the pump, air again dissolves in

the solvent. The solvent is filtered through a stainless steel or sintered

glass filter to remove any solid contaminants. Depending on the type of

solvent programmer that is employed, the supply from each reservoir

may pass either to a pump or to a valved blending device. Solvent

reservoirs are not usually thermostatted but, when necessary, the solvent

can be brought to the column temperature by the use of an appropriate

heat exchanger. The solvent containers are often situated in an enclosure

that protects the user from toxic solvent vapors such as chloroform or

aromatic hydrocarbons. Such enclosures also isolate the solvents from

atmospheric moisture.

The Gradient Programmer and the LC Pump

The Gradient Programmer

There are two basic types of solvent programmer. In the first, the solvent

mixing occurs at high pressure and in the second the solvents are

premixed at low pressure and then passed to the pump. The high-

pressure programmer is the simplest but most expensive as each solvent

requires its own pump. Theoretically, there can be any number of

solvents involved in a mobile phase program, however, most LC

analyses require only two solvents, nevertheless, up to four solvents can

be accommodated. The layout of a high-pressure gradient system is

shown in figure 2 and includes, as an example, provision for three

solvents to be mixed by appropriate programming.

Solvent passes from each reservoir directly to a pump and then to a

mixing manifold from which it passes to the sample valve and column.

The pumps control the actual program and are usually driven by

stepping motors. The volume delivery of each solvent is controlled by



5

the speed of the respective pump, which is precisely determined by the

frequency of its power supply.

Solvent

Reservoir 1

Solvent

Reservoir 2

Solvent

Reservoir 3

Pump 1

Pump 2

Pump 3

To Sample Valve

Mixing Manifold

Programme r

Figure 2. The High Pressure Gradient Programmer

The controlling frequency can be generated either by external oscillators

or, if the chromatograph is computer controlled, directly from the

computer itself.

In a low pressure programmer, the solvent from each reservoir passes to

an oscillating valve, the output from which is connected to a mixing

manifold. A diagram of the layout of a typical low-pressure solvent

programmer is shown in figure 3.

The manifold receives and mixes solvents from each of the programmed

valves. The valves are electrically operated and programmed to open and

close for different periods of time by adjusting the frequency and wave-

form of the supply. Thus, a pre-determined amount of each solvent is

allowed to flow into the manifold. The valves can also be driven either

by oscillators contained in a separate electronic programmer or by the



6

chromatograph computer which modifies the wave-form and frequency

to control the flow of each solvent.

Solvent

Reservoir 1

Solvent

Reservoir 2

Solvent

Reservoir 3

Pump

To Sample Valve

Mixing Manifold

Programmer

Valve 3

Valve 2

Valve 1

Figure 3. A Low Pressure Solvent Programmer

The LC Pump

There are a number of different types of pumps that can provide the

necessary pressures and flow-rates required by the modern liquid

chromatograph. In the early years of the LC renaissance, there were two

types of pump in common use; they were the pneumatic pump, where

the necessary high pressures were achieved by pneumatic amplification,

and the syringe pump, which was simply a large, strongly constructed

syringe with a plunger that was driven by a motor. Today the majority of

modern chromatographs are fitted with reciprocating pumps fitted with

either pistons or diaphragms.

The Pneumatic Pump

The pneumatic pump has a much larger flow capacity than the piston

type pumps but, nowadays, is largely used for column packing and not



7

for general analysis. The pneumatic pump can provide extremely high

pressures and is relatively inexpensive, but the high-pressure models are

a little cumbersome and, at high flow rates, can consume considerable

quantities of compressed air. A diagram of a pneumatic pump is shown

in figure 4.

Air

Air

Non-Re turn Valve

Non-Re turn Valve

To Sample Valve

From S olvent Reservoir

x

y

(1)

(2)

Figure 4. A Diagram of the Pneumatic Pump

It is seen in figure 4 that the total air pressure on the piston, diameter (y),

is transferred to a piston controlling the liquid pressure, of diameter (x).

Because the radii of the pistons differ, there will be a net pressure

amplification of y2

x2. For example, if the upper piston is 5 cm in

diameter and the lower piston 1 cm in diameter, then the amplification

factor will be 52

12 25. It is seen that the system can provide very high

pressures in a relatively simple manner.

The pump operates in the following manner. Air enters port (1) and

applies a pressure to the upper piston that is directly transmitted to the



8

lower piston. If connected to a liquid chromatograph, or the column

packing manifold, liquid will flow out of the left hand side non-return

valve as shown in the diagram which continues until the maximum

movement of the piston is reached. At the extreme of movement a micro

switch is activated and the air pressure transferred from port (1) to port

(2).

The piston now moves in the opposite direction and draws mobile phase

through the right hand non-return valve refilling the cylinder with

solvent. Again on reaching the limit of the piston movement a second

micro switches the air supply from port (2) back to port (1) and the

process repeated.

The refilling process is extremely fast and, if an appropriate pulse

dampener is used, the outlet pressure remains reasonably constant during

the refill cycle.

Non-Return Valves

For efficient function, it is important that while the piston compresses

the solvent to express it from the exit port, the flow is completely

stopped at the inlet port. Conversely, when the pump draws fresh solvent

into the cylinder during refill, the non-return valves must allow solvent

to flow through the inlet valve but, flow-back from the exit valve must

be completely stopped. This is achieved by the use of efficient non-

return valves. Most non-return valves are of similar design and the

construction of a typical valve is shown in figure 5.

The critical part of the valve consists of a synthetic sapphire ball resting

on a seat. The seat may be of stainless steel, captured PTFE or, more

commonly, also of sapphire. When the flow is directed against the ball

the ball moves forward allowing the liquid to flow past it. When the

direction of pressure changes resulting in potential flow-back through

the valve, the ball is driven back onto its seat and stops the flow.



9

With careful design and exacting construction these types of valve can

be extremely efficient. In practice, to ensure the most effective

performance, a single non-return valve assembly usually contains two

non-return ball valves connected in series as shown in figure 5.

Ball Valves

Flow Di recti on

Flow Impede d

Valve Se atings

Figure 5. The Design of a Typical Non-Return Valve

The Syringe Pump

The syringe pump is a large, electrically operated simulation of a

hypodermic syringe. Although used in the early days of LC renaissance,

it is rarely used today as, due to its design, it can provide only a limited

pressure and the volume of mobile phase available for use is restricted to

the pump volume. Unless the separation is stopped while the pump is

refilled and the development subsequently continued, the pump can only

elute solutes that have retention volumes equal or less than the pump

capacity. A diagram of a syringe pump is shown in figure 6.



10

The pump takes the form of a large metal syringe, the piston being

propelled by an electric motor and driven by a worm gear. The speed of

the motor determines the pump delivery. Another motor actuates the

piston by a different system of gearing to refill the syringe rapidly when

required. The solvent is sucked into the cylinder through a hole in the

center of the piston and between the piston and the outlet there is a coil

that acts as a dampener.

Piston Drive Motor

Fast Re fill Motor

To Col umn

Mobile Phase Reservoir

Drive Control Manual Rewind

Courtesy of the Perkin Elmer Corporation

Figure 6. The Syringe Pump

This type of pump is still occasionally used for the mobile phase supply

to microbore columns that require small volumes of mobile phase to

develop the separation. It is also sometimes used for reagent delivery in

post column derivatization as it can be made to deliver a very constant

reagent supply at very low flow rates.

The Single Piston Reciprocating Pump

The single piston reciprocating pump was the first of its type to be used

with high efficiency LC columns (columns packed with small particles)



11

and is still very popular today. It is simple in design and relatively

inexpensive. A diagram of the single piston pump is shown in figure 7.

Most pistons of modern LC pumps are made of synthetic sapphire to

reduce wear and extend the working life of the pump. The cylinder is

usually made of stainless steel and is attached to two non-return valves

in line with the inlet and outlet connections to the pump.


Cylinde r

Cam Driving the Piston

Driving Motor

Sapphi re Pi ston

Solvent to Column System

Non-Re turn Valve

Non-Re turn Valve

Re turn Spring

Figure 7. A Single Piston Reciprocating Pump

The piston is driven by a stainless steel cam, which forces the piston into

the cylinder expressing the solvent through the exit non-return valve.

After reaching the maximum movement, the piston follows the cam and

returns as a result of the pressure exerted by the return spring. During

this movement the cylinder is loaded with more solvent through the inlet

non-return valve. The shape of the cam is cut to provide a linear

movement of the piston during expression of the solvent but a sudden

return movement on the refill stroke. In this way the pulse effect that

results from the refill action is reduced.



12

The pulses, however, are not completely eliminated and the detector

noise resulting from these pulses is probably the most serious

disadvantage of the single piston pump. Nevertheless, as a result of its

low cost it remains one of the more popular LC pumps.

The Rapid Refill Pump

In order to avoid the refill pulses resulting from a single piston pump, a

number of rapid refill systems have been developed. The designs have

ranged from cleverly designed actuating cams to drive the piston rapidly

in the refill mode to electronically operated piston movements. One

successful approach to this problem is exemplified by the pump design

shown in figure 8.

To Column

To Column

1 ml pe r mi nute

De livery Pi ston Solvent Supply

Pis ton

From S olve nt Supply System

De livery Pi ston

Solvent Supply Pis ton

100 ml per minute

From S olve nt Supply System

Courtesy of Perkin Elmer Inc.

Figure 8. The Rapid Refill Pump



13

The pump consists of two cylinders and a single common piston. The

expression of the solvent to the column is shown in the upper part of

figure 8. As the piston progresses to the right, solvent is displaced to the

column system and, simultaneously, fresh solvent is withdrawn from the

solvent reservoir into the right hand chamber. When the piston arrives at

the extent of its travel, a step in the driving cam is reached and the

piston is very rapidly reversed. As a result the contents of the chamber

on the right-hand side are displaced into the left-hand chamber. This

situation is shown in the lower part of figure 8. The transfer rate of the

solvent to the left-hand chamber is 100 times faster than the delivery rate

to the column and consequently reduces the refill-pulse very

significantly. In addition, if a solvent gradient is being used and the

right-hand chamber is being filled with a solvent mixture, excellent

mixing is achieved during the refill of the left-hand chamber. It is clear,

however, that there will not be a smooth transition from one solvent

concentration to the next but will be a step-wise change.

An alternative approach to the elimination or reduction of pump pulses

and one which is probably the more successful (though more expensive)

is the use of twin pump heads. During the operation of a two-headed

pump, one cylinder is filled while the other is delivering solvent to the

column.

The Twin-Headed Pump.

The cylinders and pistons of a two-headed pump are constructed in the

same manner as the single piston pump with sapphire pistons and

stainless steel cylinders fitted with non-return valves to both the inlet

and outlet. The driving cams of both pistons are carefully cut to provide

an increase in flow from one pump while the other pump is being filled.

This compensate for the loss of delivery during the refill process and the

consequent fall in pressure. A diagram of a twin-headed pump is shown

in figure 9. It is seen that there is a common supply of mobile phase

from the solvent reservoir or solvent programmer to both pumps and the

output of each pump joins and the solvent then passes to the sample

valve and then to the column. In the diagram, a single cam drives both



14

pistons, but in practice, to minimize pressure pulses, each pump usually

has its own cam drive from the motor.

The cams are carefully designed to produce a virtually pulse-free flow.

The displacement volume of each pump can vary from 20 or 30

microliters to over one milliliter but the usual displacement volume for

the typical pump is about 250 l.


Driving Motor

Sapphire Pis ton

Solvent to Column System

Non- Re turn Valve

Re turn Spring

Re turn Spring

Cylinde r

Cam Driving the Piston

Non-Re turn Valve

Sapphire Pis ton

Cylinde r

Non-Re turn Valve

Non-Re turn Valve

Figure 9. The Twin-Headed Pump

A stepping motor drives the pump and thus the delivery depends on the

frequency of the supply fed to the motor. As a consequence, the pump

can have a very wide flow rate range from a few microlitres per minute

to over 10 ml per minute.

The Diaphragm Pump



15

The unique property of the reciprocating diaphragm pump is that the

actuating piston does not come into direct contact with the mobile phase

and thus, the demands on the piston-cylinder seal are not so great. The

diaphragm has a relatively high surface area and thus, the movement of

the diaphragm is relatively small and consequently the pump can be

operated at a fairly high frequency. High frequency pumping results in a

very significant reduction in pulse amplitude and, in addition, high

frequency pulses are more readily damped by the column system.

Nevertheless, it must be emphasized that diaphragm pumps are not

pulseless. A diagram of a diaphragm pump, showing its mode of action

is depicted in figure 10

Solvent from Reservoir

Solvent from Reservoir

To Column

To Column

2

34

1

Figure 12. The Action of a Diaphragm Pump

The wheel driving the crank rotates in an anti-clockwise direction and in

position (1) the diaphragm has been withdrawn causing the pumping

cavity behind the diaphragm to be filled with solvent. In position (2), the

piston advances and when it passes the pumping fluid inlet, it starts



16

compressing the diaphragm expressing solvent to the column. In

position (3) the diaphragm has been compressed to its limit and the

piston starts to return. In position (4) the piston moves back

withdrawing the diaphragm sucking liquid into the pumping cavity

ready for the next thrust. The inlet from the solvent supply and the outlet

to the column are fitted with non-return valves in the usual manner. Due

to the large volume of the pumping cavity, any gradient profile would be

seriously distorted so this type of pump is not often used for analytical

purposes but is often used in preparative chromatography.

The Sample Valve

In general, LC sample valves must be able to sustain pressures up to

10,000 p.s.i., although they are most likely to operate on a continuous

basis, at pressures of 3,000 p.s.i. or less. The higher the operating

pressure the tighter the valve seating surfaces must be forced together to

eliminate any leak. It follows that any abrasive material, however fine,

that passes into the valve can cause the valve seating to become scored

each time it is rotated which will ultimately lead to leaks. This will cause

the sample size to vary between samples and eventually affect the

accuracy of the analysis. It follows that any solid material must be

carefully removed from any sample before filling the valve. The two

basic types of LC sample valve have been discussed in Book 1. In LC

however, there is a third type of valve which is similar to the external

loop valve but contains an extra loading port and behaves like an

internal loop valve. a diagram of which is shown in figure 13.



17

Mobile Phase

Mobile Phase

Sample LoopSample Loop

To ColumnTo Column

Sample In Sample In

To WasteTo Waste

Position A Position B

Ve ntVe nt

Ne edle Port Ne edle Port

Courtesy of Rheodyne Instruments Inc.

Figure 13. The Modified External Loop Sample Valve

The basic difference between this type of valve and the normal external

loop sample valve is the introduction of an extra port at the front of the

valve. This port allows the injection of a sample by a syringe directly

into the front of the sample loop. Position (A) shows the inject position.

Injection in the front port causes the sample to flow into the sample

loop. The tip of the needle passes through the rotor seal and, on

injection, is in direct contact with the ceramic stator face. Note the

needle is chosen so that its diameter is too great to enter the hole.



18

Valve Body Rotor

Pre-load Assembly

Connecting Slot

Courtesy of Valco Instruments Inc.

Figure 14. A Simple Form of the LC Sample Valve

After injection, the valve is rotated to position (B) and the mobile phase

flushes the sample directly onto the column. The sample is actually

forced out of the beginning of the loop so it does not have to flow

through the entire length of the loop. This type of injection system is

ideally suited for quantitative LC, and is probably by far the most

popular injection system in use. Sample valves based on this design are

available from a number of manufacturers.

As a consequence of the high pressures that must be tolerated, LC

sample valves are usually made from stainless steel. The exception to the

use of stainless steel will arise in biochemical applications where the

materials of construction may need to be biocompatible. In such cases

the valves may be made from titanium or some other appropriate

biocompatible material. It should be stressed that only those surfaces

that actually come in contact with the sample need to be biocompatible

and the major parts of the valve can still be manufactured from stainless

steel. The actual structure of the valve varies a little from one

manufacturer to another but all are modifications of the basic sample

valve shown in figure 14.



19

The valve usually consists of five parts. Firstly there is the control knob

or handle that allows the valve selector to be rotated and thus determines

the load and sample positions. Secondly, a connecting device that

communicates the rotary movement to the rotor. Thirdly the valve body

that contains the different ports necessary to provide connections to the

mobile phase supply, the column, the sample loop if one is available, the

sample injection port and finally a port to waste. Then there is the rotor

that actually selects the mode of operation of the valve and contains slots

that can connect the alternate ports in the valve body to provide loading

and sampling functions. Finally there is a pre-load assembly that

furnishes an adequate pressure between the faces of the rotor and the

valve body to ensure a leak tight seal.

Column Switching

The technique of column switching can increase the versatility of the

liquid chromatograph significantly. An example of a six-port valve

arranged for column switching is given in figure 15. The arrangement

incorporates the same valve as that used for the external loop sampling

system. It is seen that column (1) is connected between ports (5) and (6)

and column (2) is connected between ports (2) and (3). Mobile phase

from a sample valve, or more usually from another column, enters port

(1) and the detector is connected to port (4). In the initial position of the

rotor shown in the diagram on the left hand side, the rotor slots connect

ports (1) and (6), (2) and (3) and (4) and (5).

This results in mobile phase passing from port (1) to port (6), through

column (1) to port (5), from port (5) to port (4) and out to the detector.

Thus, the separation will take place in column (1). The ports connected

to column (2) are themselves connected by the third slot and thus

isolated.



20

Column 1

Column 1

Column 2

Column 2

Mobile Phase

Mobile Phase

To Detector To Detector

112 2

3344

5 5

6 6

Courtesy of Valco Instruments Inc.

Figure 16. Valve Arrangement for Column Switching

When the valve is rotated, the situation is depicted on the right hand side

of figure (7); port (1) is connected to port (2), port (3) connected to port

(4) and port (5) connected to port (6). This results in the mobile phase

from either a sample valve or another column entering port (1) passing

to port (2) through column (2) to port (3), then to port (4) and then to the

detector. The ports (5) and (6) are connected, this time isolating column

(1). This arrangement allows either one of two columns to be selected

for an analysis or part of the eluent from another column pass to column

(1) for separation and the rest passed to column (2). This system,

although increasing the complexity of the column system renders the

chromatographic process far more versatile. The number applications

that require such a complex chromatographic arrangement are relatively

small, nevertheless, when required, column switching can provide a

simple solution to certain difficult separation problems.

Column Ovens

The effect of temperature on LC separations is often not nearly so

profound as its effect in GC separations, but can be critical when closely

similar substances are being separated. In LC a change in temperature

will change the free energy of the solute in both phases, (generally in a



21

commensurate manner) and so the net change in the free energy

difference with temperature, which controls the magnitude of the

absolute retention, can be relatively small. Its effect on relative

retention, however, can be very significant and, in fact, be the

determining factor in achieving a satisfactory resolution. (5-7) The

effect of temperature on diffusivity will be similar in both GC and LC. An

increase in temperature will increase the diffusivity of the solute in both

phases and thus increase the dispersion due to longitudinal diffusion

and decrease dispersion due to resistance to mass transfer. As a result, at

the optimum velocity, the efficiency of both the LC and GC column will

be largely independent of temperature, however, the optimum velocity

will be higher at higher temperatures and provide the potential for faster

analyses. Due to the lesser effect of temperature on solute retention in

LC (compared to that in GC), temperature is not nearly so critical in

governing absolute retention time but is often essential in achieving

adequate resolution, particularly between closely eluting solutes such as

isomers. In contrast to the GC column, the thermal capacity of an LC

column is much higher as the specific heats of liquids are much greater

than those of a gas. As a consequence, a high heat capacity

thermostatting fluid is necessary and if retention measurements need to

be precise, air ovens would not be ideal for thermostatting LC columns.

On the other hand, liquid thermostatting media are rather messy to use

and tend to make column changing difficult and lengthy. However, if

accurate data is required, good temperature control may be essential. If

precise retention measurements are not required, an air thermostatting

oven might be a reasonable compromise.

Detectors

A large number of LC detectors have been developed over the past thirty

years based on a variety of different sensing principles. However, only

about twelve of them can be used effectively for LC analyses and, of

those twelve, only four are in common use. The four dominant detectors

used in LC analysis are the UV detector (fixed and variable wavelength)

the electrical conductivity detector, the fluorescence detector and the

refractive index detector. These detectors are employed in over 95% of



22

all LC analytical applications. These four detectors will be described and

for those readers requiring more information on detectors are referred to

Book 5 on Liquid Chromatography Detectors. The subject of detector

specifications will not be discussed here but will also be dealt with in

detail in Books 4 and 5. Detector sensitivities and detector linearity will,

however, be given for each of the four detectors.

The UV Detector

The UV detector is by far the most popular and useful LC detector that

is available to the analyst at this time. This is particularly true if multi-

wavelength technology is included in this class of detectors. Although

the UV detector has some definite limitations (particularly for the

detection of non polar solutes that do not possess a UV chromaphores) it

has the best combination of sensitivity, linearity, versatility and

reliability of all the LC detectors so far developed .

Most compounds adsorb UV light in the range of 200-350Å including

all substances having one or more double bonds ( electrons) and all

substances that have unshared (non bonded) electrons; e.g. all olefins, all

aromatics and all substances containing >CO, >CS, -N=O and

N N groups. The relationship between the intensity of UV light

transmitted through the detector cell and solute concentration is given by

"Beers' Law,

IT Io ekLc

where, (Io) is the intensity of the light entering the cell,

(IT) is the intensity of the transmitted light,

(L) is the path length of the cell,

(c) is the concentration of the solute,

(k) is the molar extinction coefficient of the solute for the

specific wavelength of the UV light.

or, Ln IT ln Io kLc



23

Rearranging,

IT Io 10k'Lc

where (k') is the molar extinction coefficient of the solute.

Differentiating,

LnIT

Io

ckL

It is seen that there are two factors that control the detector sensitivity,

the magnitude of the extinction coefficient of the solute being detected

(which will depend on the wavelength of the UV light that is used) and

the path length of the light passing through the cell. Thus, although the

minimum detectable concentration can be changed by selecting a light

source of different wavelength, the cell length can not be increased

indefinitely to provide higher sensitivity as long cells will provide

excessive peak dispersion with consequent loss of column resolution. It

follows, that the optimum detector cell design involves the determination

of the cell length that will provide the maximum sensitivity and at the

same time constrain detector dispersion to a minimum so that there is

minimum loss in resolution.

The Fixed Wavelength Detector

There are two types of UV detector the fixed wavelength detector and

the multi-wavelength detector. A diagram of a Fixed Wavelength UV

Detector is shown in figure 17.

The detector consists of a small cylindrical cell (2.0 to 10.0 l in volume)

through which flows the eluent from the column. UV light from an

appropriate UV lamp passes through the cell and falls on a UV photo

electric cell. In the fixed wavelength detector the wavelength of the light

depends on the type of lamp that is used. There are a number of lamps



24

available the at provide of wavelengths ranging from about 210 nm to

280 nm.

UV Lamp

Quartz Windows

Quartz Windows

Sample Cell

Reference Cell

Lens

Photo Cells

Figure 17. The Fixed Wavelength UV Detector

The common lamps that are commercially available at the time of

writing this book are as follows:-

Lamp Type Emission Wavelengths

Mercury Vapor Lamp 253.7 nm

Zinc Vapor Lamp 2123.9 nm and 307.6 nm

Cadmium Vapor Lamp 228.8, 326.1,340.3, and 346.6 nm

The mercury vapor lamp is the most popular as it has an emission

wavelength that allows the detector to sense a wide range of solute

types. The detector usually contains both a sample and reference cell and

the output from the reference cell is compared to that from the sample

cell The difference is fed to a non linear amplifier that converts the

signal to one that is linearly related to concentration of solute in the

sample cell. The fixed wavelength detector is the least expensive and, as

the majority of the light is emitted at a specific wavelength(s) it has a



25

high intensity, and thus, a higher intrinsic sensitivity than the multi-

wavelength UV detectors. However, the multi-wavelength detector can

often compensate for the lower sensitivity by choosing a wavelength that

has the highest extinction coefficient for the solutes of interest. Average

specifications for commercially available fixed wavelength UV detectors

are as follows:-

Typical Specifications for a Fixed Wavelength UV Detector

Sensitivity (toluene) 5x 10-8 g/ml

Linear Dynamic Range 5 x 10-8 to 5 x 10-4 g/ml

Response Index 0.98 - 1.02

Figure 18. The High Speed Separation of a Two Component

Mixture

By the use of very small sensing cells and electronic systems with very

small time constants, the fixed wavelength detector can be designed to

give a very fast response at high sensitivity and very low dispersion and

for this reason it can be used for high speed separations. An example of

the rapid separation of a two component mixture (8) is shown in figure

18. The reason for separating benzene and benzyl alcohol in 2.6

seconds remains (to say the least) obscure and figure 1 is obviously and



26

example of "Chromatography Show Biz". Nevertheless, it does

demonstrate that columns can be designed and detectors developed that

can provide extremely fast analyses.

The Multi-Wavelength Detectors

Multi-Wavelength UV detectors utilize a single (perhaps more

accurately a narrow range) of wavelengths to detect the solute. Most

multi wavelength UV detectors can also provide a UV spectrum of the

eluted solute if appropriately arranged. There are two types of multi-

wavelength detectors the dispersion detector that monitors the eluent at

one wavelength only and the diode array detector that monitors the

eluted solute over a range of wavelengths simultaneously. The former

passes the light from a broad emission light source through a

monochrometer, selects a specific wavelength and allows it to pass

through the detecting cell. The second, also uses a broad emission light

source, but all the light is allowed to pass through the sensing cell and

subsequently the light is dispersed by means of a holographic grating

and the dispersed light allowed to fall on an array of diodes.

The Multi-Wavelength Dispersive Detector

A diagram of the Multi-Wavelength Dispersive Detector is shown in

figure 19.

Photo CellLens

Flow

Cel lLens

Plane

Mi rror

Curved

Mi rror

Curved

Mi rror

MONOC HROMATOR

Grating

Curved Mirror

Deuterium

Lamp



27

Figure 19. The Multiwave Length Dispersive Detector

Light from a broad wavelength source such as a deuterium or xenon

discharge lamp is collimated by two curved mirrors onto a holographic

diffraction grating. The dispersed light is focused by means of a curved

mirror, on to a plane mirror and light of specific wavelength selected by

appropriately positioning the angle of the plane mirror. Light of the

selected wavelength is then focused by means of a lens through the flow

cell and consequently, through the column eluent. The exit beam from

the cell is focused by another lens onto a photocell which gives a

response that is some function of the intensity of the transmitted light.

The detector is usually fitted with a scanning facility that, by arresting

the flow of mobile phase, allows the spectrum of the solute contained in

the cell to be obtained.

Due to the limited information provided by UV spectra and the

similarity between many spectra of widely different types of compound,

UV spectra are not very reliable for solute identification. The technique

is useful, however, for determining the homogeneity of a peak by

obtaining spectra from a sample on both sides of the peak. The

technique is to normalize both spectra, then either subtract one, from the

other, and show that the difference is close to zero or take the ratio and

show it is constant throughout the peak.

A more common use of the multi-wavelength detector is to select a

wavelength that is characteristically absorbed by a particular component

or components of a mixture. This can be done to either enhance the

sensitivity of the detector to those particular solutes, or render the

detector more specific and consequently, not give a significant response

to other substances in the mixture.

The multi-wavelength dispersive detector more useful than the fixed

wavelength type of UV detector providing adequate sensitivity,

versatility and a linear response. It is however somewhat bulky, due to

the need for an relatively large internal 'optical bench', has mechanically

operated wavelength selection and requires a stop/flow procedure to



28

obtain spectra "on-the-fly". The Diode Array Detector has the same

advantages but none of the disadvantages, though, as one might expect,

is somewhat more expensive.

The Diode Array Detector

The diode array detector, although offering detection over a range of UV

wavelength, functions in an entirely different way from that of the

dispersive instrument. A diagram of a diode array detector is shown in

figure 20. Deuterium

Lamp

Achromatic Lens System

Shutter

Flow Cell

Inlet

Outlet

Photo Diode Array

Holographic Grat ing

Figure 20. The Diode Array Detector

Light from a broad emission source such as a deuterium lamp is

collimated by an achromatic lens system so that the total light passes

through the detector cell onto a holographic grating. In this way the

sample is subjected to light of all wavelengths generated by the lamp.

The dispersed light from the grating is allowed to fall on to a diode

array. The array may contain many hundreds of diodes and the output

from each diode is regularly sampled by a computer and stored on a hard

disc. At the end of the run, the output from any diode can be selected



29

and a chromatogram produced employing the UV wavelength that was

falling on that particular diode. Most instruments will permit the

monitoring of a least one diode in real time so that the chromatogram

can be followed as the separation develops. This system is ideal in that

by noting the time of a particular peak, a spectrum of the solute can be

obtained by recalling from memory the output of all the diodes at that

particular time. This gives directly the spectrum of the solute, i.e., a

curve relating adsorption against wavelength.

The performance of both types of multi-wavelength detectors are very

similar and typical values for their more important specifications are as

follows.

Typical Specifications for a Multi-Wavelength UV Detector

Sensitivity 1 x 10-7g/ml



The Electrical Conductivity Detector

The electrical conductivity detector can only detect those substances that

ionize and consequently, is largely used in the analysis of inorganic

acids, bases and salts. It has also found particular use in the detection of

organic acids and bases that are frequently required in environmental

studies and in biotechnology applications. The sensor is the simplest of

all the detectors consisting of only two electrodes situated in a suitable

flow cell.

An example of an electrical conductivity sensing cell is shown in figure

21. It consists of two electrodes situated in a suitable flow cell as

depicted in the upper diagram. The electrodes are arranged to constitute

one arm of a Wheatstone Bridge. When ions enter the detector cell, the

electrical resistance changes and the out of balance signal is fed to a

suitable amplifier. The output from the amplifier is either digitized, and



30

the binary number sent to a computer for storage, or the output is passed

directly to a potentiometric recorder. The detector actually measures the

electrical resistance between the electrodes, which by suitable non-linear

amplification, can be made to provide an output that is linearly related to

solute concentration. It is essential that an AC voltage is used across the

electrodes to measure the cell impedance to avoid electrode polarization.

The frequency of the AC potential across the electrodes is usually

around 10 kHz

Electrodes

Inlet Outlet

Conduct ivity Cell

Amplifier

Inlet Outlet

AmplifierPT FE Insulation

Figure 21. An Electrical Conductivity Detector Sensing Cell

A more practical system shown in the lower part of the diagram consists

of short lengths of stainless steel tube insulated from each other by a

PTFE connecting sleeves. For convenience, the first tube (that connected

to the column) is usually grounded (earthed). The resistance between the

inlet tube and the center tube is continuously monitored which will

constitute the resistance across the tiny gap between the tubes contained

in the first PTFE sleeve. The volume of eluent in this gap can be

extremely small and thus, the peak dispersion can also be made very

small. The resistance of the solution situated between the tubes is

inversely proportional to the electric conductivity of the solution, which

in turn, is related to the ion concentration in mobile phase. Some typical

specifications for an electrical conductivity detector are as follows.



31

Typical Specifications for an Electrical Conductivity Detector

Sensitivity 5x 10-9 g/ml



The separation of a mixture of alkali and alkaline earth cations at levels

of a few parts per million, shown in figure 22 gives an example of the

use of the electrical conductivity detector. The cations lithium, sodium,

ammonium, potassium, magnesium and calcium were present in the

original mixture at concentrations of 1, 4, 10, 10, 5 and 10 ppm

respectively.

Courtesy of Dionex Inc

Figure 22. Determination of Alkali and Alkaline Earth Cations

The column used was the IonPacCS12 (a proprietary ion exchange

column) with a mobile phase consisting of a 20 nM methanesulphonic

acid solution in water at a flow rate of 1 ml/min. The sample volume

was 25 l and the separation is also an interesting example of the use of

the ion suppression technique. If the methanesulphonic acid solution



32

was passed directly from the column through the detector it would have

a high electrical conductivity and give a large detector base current

which would swamp the signal from the ions being monitored. Thus,

subsequent to the mobile phase leaving the column (and after the

methane sulphonic acid has achieved its purpose in producing the

desired separation) the reagent must be removed to ensure that the

mobile phase entering the detector only contains those ions of interest

and minimal background conductivity. The methane sulphonic acid was

removed by passing the mobile phase through a short reverse phase

column before it entered the detector. The reverse phase will remove all

organic material by adsorption due to the strong dispersive forces that

will occur between the hydrocarbon chains of the reverse phase and the

methyl group of the methanesulphonic acid. The ion suppression column

eventually saturates and require regeneration by desorbing the methane

sulphonic acid with a strong dispersive solvent that is miscible with

water such as acetonitrile. This technique of ion suppression is

frequently used in ion exchange chromatography when using the

electrical conductivity detector. A wide variety of different types of ion

suppression columns are available but it should be pointed out that, any

suppresser system introduced between the column and the detector, will

cause some degree band spreading and consequently reduce the

resolving power of the system. It follows, that the connecting tubes and

suppression column itself must be very carefully designed to eliminate

or reduce this dispersion to an absolute minimum.

The Fluorescence Detector

The fluorescence detector is one of the most sensitive LC detectors and

for this reason is often used for trace analysis. Unfortunately, although

the detector is very sensitive, its response is only linear over a relatively

limited concentration range. In fact, the response of the detector can

only be assumed to be linear over a concentration range of two orders of

magnitude. Unfortunately, the majority of substances do not naturally

fluoresce which is a serious disadvantage to this type of detector. It

follows, that in many instances fluorescent derivatives must be

synthesized to render the substances of interest detectable. There are a



33

number of regents that have been developed specifically for this purpose

but derivatizing procedures will be discussed in detail in Book 18. A

diagram of the Fluorescence Detector is shown in figure 23. In its

simplest form, light from a fixed wavelength UV lamp passes through a

cell, through which the column eluent flows and acts as the excitation

source. Any fluorescent light that is emitted is sensed by a photo electric

cell positioned normal to the direction of exciting UV light.

Fluorescent Light

Outlet

Inlet

Fluorescent Cell

UV Lamp

UV Light

Photo Cell

Figure 23. The Fluorescence Detector

The photocell senses fluorescent light of all wavelengths but the

wavelength of the excitation light can only be changed by use of an

alternative lamp. This simple type of fluorescence detector was the first

to be developed; it is relatively inexpensive and for certain compounds

can be extremely sensitive. Typical specifications for a fluorescence

detector are as follows:-

Typical Specifications for a Fluorescence Detector

Sensitivity (Anthracene) 1x 10-9 g/ml





34

A more elaborate form of fluorescence detector uses a monochromator

to select the excitation wavelength and a second monochromator to

select the wavelength of the fluorescent light. This instrument gives the

maximum versatility and allows the maximum sensitivity to be realized

for any type of solute. The system can also provide a fluorescence

spectra by arresting the flow of mobile phase when the solute resides in

the detecting cell and scanning the fluorescent light.

The Refractive Index Detector

The refractive index detector is one of the least sensitive LC detectors. It

is very sensitive to changes in ambient temperature, pressure changes,

flow-rate changes and cannot be used for gradient elution. Despite these

many disadvantages, this detector is extremely useful for detecting those

compounds that are nonionic, do not adsorb in the UV, and do not

fluoresce. There are many optical systems used in refractive index

detectors (9) but one of the most common is the differential refractive

index detector shown diagrammatically in figure 24.

Light SourceMask

Le ns

Sample

Re ference

Amplifi er and

Power S uppl y Zero Adjust

Se nsor

Mi rror

Re corde r

Re ference

Sample

Courtesy of the Waters Chromatography



35

Figure 24. The Refractive Index Detector

The refractometer shown in figure 24 monitors the deflection of a light

beam caused by the refractive index difference between the contents of

the sample cell and that of the reference cell. A beam of light (usually

from an incandescent lamp) passes through an optical mask that

confines the beam to the region of the cell. The lens collimates the light

beam, which then passes through both the sample and reference cells to

a plane mirror. The mirror reflects the beam back through the sample

and reference cells to a lens, which focuses it onto a photo cell. The

location of the beam, rather than its intensity, is determined by the

angular deflection of the beam resulting from the refractive index

difference between the contents of the two cells. As the beam changes

its position of focus on the photoelectric cell, the output changes and the

difference signal is electronically modified to provide a signal

proportional to the concentration of solute in the sample cell.

The refractive index detector can often be a 'choice of last resort' and is

selected for those applications where, for one reason or another, all other

detectors are inappropriate or impractical. However, the detector has one

particular and unique area of application and that is in the separation

and analysis of polymers. For those polymers that contain more than ten

monomer units, the refractive index is directly proportional to the

concentration of the polymer and is practically independent of the

molecular weight. A quantitative analysis of a polymer mixture can,

therefore, be obtained by the simple normalization of the peak areas in

the chromatogram (there being no need for the use of individual

response factors). Some typical specifications for the refractive index

detector are as follows:-

Typical Specifications for a Refractive Index Detector

Sensitivity (benzene) 1x 10-6 g/ml





36

A typical application of the RI detector is for carbohydrate analysis.

Carbohydrates do not adsorb in the UV, do not ionize and although

fluorescent derivatives can be made, the procedure is tedious and time

consuming. An example of such an application is shown in figure 25 by

the separation of the products of cyclodextrin hydrolysis.

Courtesy of TOYO SODA Manufacturing Co. LTD

Figure 25. The Separation of Hydrolyzed cyclodextrin

A TSK gel G-Oligo-PW column 7.8 mm I.D.. and 30 cm long was used

for the separation which was carried out at 60˚C and a flow rate of 1

ml/min. The TSK gel packing is a vinyl polymer based material suitable

for separation by size exclusion using aqueous solvents. It is seen that

the products of the hydrolysis are well separated and almost all of the

oligomers are resolved.

The Tridet Multi Functional Detector



37

The popularity of the UV detector, the electrical conductivity detector

and the fluorescence detector motivated Schmidt and Scott (10,11) to

develop a trifunctional detector that detected solutes by all three method

simultaneously in a single low volume cell. A diagram of their detector

is shown in figure 26. The UV adsorption system consists of a low-

pressure mercury lamp (major emission at 254 nm) and a solid state

photo cell with quartz windows so it responds to light in the UV region.

The cell is 3 mm long and is terminated at on end by a cylindrical quartz

window and at the other by a quartz lens.

UV Lamp

Fluorescence Photo Cell UV Adsorpti on

Photo Cell

Quartz Lens

Quartz Lens

Flow Ce ll

Inl et

Outlet

Conduit Tubes also Se rving as Electrodes

Re taini ng Springs

Figure 26. The Trifunctional Detector

The lens focuses the transmitted light on to the photocell. There are two

stainless steel discs separated by a 1 cm length of Pyrex tube adjacent to

the quartz windows. Mobile phase enters and leaves the detector cell

through radial holes in the periphery of the stainless steel discs. The steel

discs act as the electrodes for conductivity detection. Normal to the

Pyrex tube is another photocell that receives fluorescent light emitted at



38

right angles to the UV excitation light. The output from each sensor is

processed by an appropriate amplifier to provide an output that is

linearly related solute concentration.

The column eluent is, thus, continuously and simultaneously monitored

by UV adsorption, fluorescence and electrical conductivity. This

detector is very versatile and is supplied with a column, sample valve,

pump and recorder. Chromatograms showing the simultaneous use of all

three detector functions are depicted in figure 27.

UV Adsorption

Function

Electrical.Conductivity

Function

Fluorescence

Function

The column used was a Pecosphere™ 3 mm in diameter and 3 cm long carrying a

C18 stationary phase. The mobile phase was a mixture of methanol(75%) and water

(25%) at a flow rate of 2 ml/min. The solutes were 1 benzene, 2 toluene, 3 ethyl

benzene, 4 isopropyl benzene, 5 t-butylbenzene, 6 anthracene, and 7 sodium

chloride.

Courtesy of Bacharach Inc.

Figure 27. Chromatograms Demonstrating the Simultaneous

Monitoring of a Mixture by all Three Detector Functions



39

It is seen that the anthracene is clearly picked out from the mixture of

aromatics by the fluorescence detector and the chloride ion, not shown

at all by the UV adsorption or fluorescence detectors, is monitored by

the electrical conductivity detector. The simultaneous use of all detector

functions make this detector very useful but, the real advantage of the

trifunctional detector is that it allows the analyst a choice of the three

most useful detector function in one detecting system. In addition, any

of the three functions can be chosen at the touch of a switch and without

any changes in hardware. An example of the use of the three individual

detector function in the analysis of three quite different types of sample

demonstrates this advantage.

Column-PecoSphere Column Pecosphere Column Pecosphere 3 Size 3 mm x 3 cm C18 Size 4.6 mm x 15 cm C18 Size 3 mm x 3 cm C18 17% Methanol/Water 90% Acetonitrile/Water 1nMtetrabutyl-

ammonium hydroxide and buffer

Flow-Rate 3.0 ml/min. Flow-Rate 2 ml/min. 1.5 ml/min.

UV Detector Fluorescence Detector Conductivity Detector

Sample Composition

1 Theobromine 1 Naphthalene 1 Solvent peak

2 Theophyline 2 Fluorene 2 Chloride ions

3 hydroxyethyltheophyline 3 Acenaphthene 3 Nitrite ions

4 Caffeine 4 Phenanthrene 4 Bromide ions

5 Anthracene 5 Nitrate ions

6 Fluoranthene 6 Phosphate ions

7 pyrene 7 Phosphite ions



40

8 Benzo(a)anthracene 8 Sulfate ions

9 Chrysene 9 Iodide ions

10 Benzo(b) Fluoranthene

11 Benzo(k) Fluoranthene

12 Benzo(k) Fluorantmene

13 Dibenz(a,h)anthracene

14 Idenol(1,2,3-cd))pyrene

15 Benzo(ghi)perylene

Figure 28. Chromatograms from the Trifunctional Detector

Data Acquisition and Processing

The data acquisition and processing system has been discussed in Book

1 and Book 2 and is dealt with in further detail in Books 4 and 5.

However, a general outline of the data system layout is included here

and is given in figure 29 in the form of a block diagram that shows the

individual steps employed.

Detector Output

Scaling Amplifier

Analogue to Digital Converter

Interface

Computer

Presentation of Results

Figure 29. A Typical Data Acquisition and Processing System

The output from the detector, usually in millivolts, is passed to a scaling

amplifier that converts the signal to a voltage that is acceptable to the



41

analog to digital (A/D) converter The A/D converter changes the voltage

output to a binary number which is temporarily stored in a register. This

process is continuously repeated at a defined rate, called the 'sampling

rate'. The current binary number, stored in the register is regularly

sampled by the computer and stored (usually on hard disk). On

completion of the analysis the computer accesses all the data from store,

calculates the retention report, compares peak heights or peak areas to

provide the quantitative analysis according to the processing program

that is used and finally prints out the results in tabulated form.

Modern data processing software often includes routines that can

process chromatograms where the components of the sample are

incompletely resolved. The routines deconvolute the individual peaks

from the composite envelope and calculate the area of the individual de-

convoluted peaks. Such algorithms can be used very effectively on peaks

that are entrained in the tail of a major peak but are not so accurate for

composite envelopes containing many unresolved peaks.

It should be emphasized that clever algorithms or subroutines are a

poor substitute for good chromatography.

The chromatographic system should, wherever possible, be optimized to

obtain complete resolution of the mixture and not place too much

reliance on mathematical techniques to aid in the analysis.

Liquid Chromatography Stationary Phases

Traditionally the stationary phase used in LC has been silica gel, which

separates solutes largely on the basis of polarity, although, due to its

unique structure, silica gel also exhibits strong exclusion characteristics.

The bonded phases were introduced to provide a material that would

separate solutes by dispersive interactions and also to provide some semi

polar stationary phases. The bonded phases were also based on silica

gel. Subsequently, polymeric stationary phases were introduced to



42

provide materials that were insoluble in water and that were stable at

extremes of pH.

Silica Gel

Silica gel is manufactured by releasing silicic acid from a strong solution

of sodium silicate by hydrochloric acid. (Sodium silicate is prepared by

heating sand at a high temperature in contact with caustic soda or

sodium carbonate).

Initially, silicic acid is released,

Na2SiO3 +H2O + 2HCl = Si(OH)4 + 2NaCl

and then the free acid quickly starts to condense with itself with the

elimination of water to form dimers, trimers and eventually polymeric

silicic acid. The polymer grows, initially forming polymer aggregates

and then polymer spheres, a few Angstrom in diameter. These polymeric

spheres are called primary silica particles. These primary particles

continue to grow until, at a particular size, the surface silanol groups on

adjacent primary polymer particles, condense with the elimination of

water. This condensation causes the primary particles to adhere to one

another and at this stage the solution begins to gel. During this process,

the primary particles of silica gel will have diameters ranging from a few

Angstrom to many thousands of Angstrom depending on the conditions

of formation.

The size of the primary particles depend, among other factors, on the

conditions of synthesis, (e.g., reaction temperature, the pH of the

mixture at the time of gelling and even the subsequent treatment of the

gel, including the conditions of washing). The product at this stage is

called a 'hydrogel'.

If a neutralized silicate solution is allowed to age with gentle stirring to

prevent gel formation, Ostwald Ripening occurs, causing the larger

primary particles to grow in size at the expense of the formation of

smaller particles. In general, the aging and ripening of silica solutions



43

will increase the size of the larger particles, which, in turn, will decrease

the surface area of the silica and increase its porosity.

It is apparent that the formation of the primary particles of condensed

silica, and their subsequent fusion by gelling, that confers on silica gel

its high porosity and high surface area, which are so important in its use

as a stationary phase in LC. Furthermore, as has already been stated, it is

the condensation of the surface silanol groups of the primary particles

that causes their adhesion and the onset of gel formation and,

consequently, the mechanical strength of the gel.

After the gel has first set, it is a very soft and is usually transferred to

vats or trays where it is then allowed to stand for a number of days.

During this period, condensation between the primary particles

continues to take place and the gel shrinks and exudes salinated water.

This shrinking process accompanied by saline elimination is called

sinerisis and eventually a firm, almost rigid gel is produced which is still

called the hydrogel. The compact hydrogel is then well washed under

controlled conditions to eliminate the last of the sodium chloride and

then heated for a few hours at 120oC. The resulting product is a hard

amorphous mass called the xerogel. The xerogel, ground and graded is

the material that is used for packing LC columns and manufacturing

bonded phases. The product, prepared in this way, is called irregular

silica gel, to differentiate it from spherical silica gel, which is prepared

by employing an entirely different synthetic procedure. Irregular silica

gel is one of the basic materials from which bonded phases can be

prepared.

The Preparation of Spherical Silica Gel

Spherical particles of silica can be prepared by spraying a neutralized

silicate solution (the colloidal silica sol.) into fine droplets before gelling

has taken place and subsequently drying the droplets in a stream of hot

air. It has also been shown possible to disperse a silica sol in the form of

an emulsion in a suitable organic solvent where the droplets gel in

spherical form (12). Unfortunately, details for the preparation of



44

spherical silica have tended to be kept very confidential for commercial

reasons and so information is a little sparse. One of the first methods

reported was that of Le Page et al (13,14). A stable silica sol. (generated

at low pH so that gelling does not take place) is passed through a non

aqueous solvent in such a manner as to produce droplets. These droplets

rapidly solidify and are then filtered off, dried, and heated to 400˚C to

800˚C to form a rigid xerogel. The structure of the resulting silica is

strongly affected by the alkali metal content and the calcining

temperature. The higher the temperature employed the lower the surface

area and pore volume of the resulting silica.

An alternative method devised and patented by Unger (15) called the

polyethoxy silane procedure involved a two-stage process. Firstly

tetraethoxy-silane is partially hydrolyzed to polyethoxysiloxane, a

viscous liquid, which is then emulsified in an ethanol water mixture by

vigorous stirring. The stirring produces spheres of polyethoxysiloxane

which by hydrolytic condensation, initiated by a catalyst, are changed to

silica hydrogel. The hydrogel spheres are then washed and converted to

the xerogel by heating. Spherical silica gel is readily available

commercially, both in the form of silica and as different types of bonded

phase. Today the majority of silica-based LC column packings are

spherical although with modern methods of grinding, the so-called

irregular particles are more rounded and the difference between the

performance of spherical and irregular particles is less clearly defined.

The Structure of Silica Gel

The matrix of the primary silica gel particle consists of a core of silicon

atoms joined together with oxygen atoms by siloxane bonds (silicon-

oxygen-silicon bonds). On the surface of each primary particle some

residual, uncondensed hydroxyl groups from the original polymeric

silicic acid remain. These residual hydroxyl groups confer upon silica

gel its polar properties. These hydroxyl groups react with the silane

reagents to form bonded phases. The silica surface is quite complex and

contains more than one type of hydroxyl group, strongly bound or

'chemically' adsorbed water and loosely bound or 'physically adsorbed'



45

water. There are three types of hydroxyl group. The first is a single

hydroxyl group attached to a silicon atom, which has three siloxane

bonds joining it to the gel matrix. The second is one of two hydroxyl

groups attached to the same silicon atom, which, in turn, is joined to the

matrix by only two siloxane bonds. These twin hydroxyl groups are

called Geminal hydroxyl groups. The third is one of three hydroxyl

groups attached to a silicon atom, which is now only joined to the silica

matrix by only a single siloxane bond. An example of each type of

hydroxyl bond is shown in Figure 30.

Employing NMR techniques Sindorf and Maciel, (16,17) has shown

that the single hydroxyl group is likely to be the most prolific. The next

most common is the geminal hydroxyl groups followed by the tertiary

hydroxyl group.

O

O

H

Si O H

O

H

O

Si

SiSiSi

Si

OO

OO

O O

O SiO

H H

O

H

O H

S i ngle S i lanol Group

Two Sil anol Groups (Geminal Groups)

Three S il anol Groups

Figure 30. Different Forms of Hydroxyl Group that can Occur on

the Surface of Silica Gel.

The silica surface, however, has additional complexities. Water can be

hydrogen bonded to the hydroxyl groups and multi-layers of water

physically adsorbed on top of these. Water can be hydrogen bonded to

the silica gel surface in a number of different ways, which are depicted

in figure 31.



46

Si Si Si Si SiSi

O O O O O

O O O O O O

H H H H H

OH

H

H H

O

H

O

H

Figure 31. Different Ways in Which Water May be Hydrogen

Bonded to Silica Gel Hydroxyl Groups

None of the above structures has been confirmed in an unambiguous

manner but all are reasonably possible. The center and right hand side

structures contain a type of double hydrogen bond and would have high

energies of formation and, thus, more stable than the simple hydrogen

bond depicted on the left. The right hand structure would be particularly

stable as it constitutes a four membered hydrogen bonded ring similar to

that, which might be expected to form in the strong association of water

with itself.

Silica gel adsorbs relatively large quantities of water, which was

explained on the basis of multi-layer adsorption. This concept was

supported by Vleeskens (18,19) and experimentally validated by

gravimetric measurements (20). An example of one type of multi-layer

adsorption is shown in figure 32.



47

SiSi

O

O O

H

OH

H

OH

H

OH

H

H

Figure 32. Multi-Layers of Physically Adsorbed Water

The multi-layer adsorption depicted in figure 32 is much over

simplified, as adsorption could also take place on the surface of siloxane

bonds as well.

Probably the most informative experimental procedure to help in the

elucidation of the structure of the silica surface is thermogravimetric

analysis (TGA). Formally, the silica was heated for known times at

known temperatures and the loss in weight noted (21). With the advent

of the programmed TGA apparatus this procedure has been simplified a

great deal and the basic thermogram can now provide considerable

information on the nature of the surface hydroxyl groups and adsorbed

water (22).

The Thermogravimetric Analysis of Silica Gel

The sample is suspended from the arm of a continuously recording

micro balance in a temperature controlled furnace and is heated from a

defined starting temperature to a specified final temperature at a

designated heating rate usually given as temperature change per unit

time. The temperature and the sample weight are continuously digitized

and the data stored. The results can then be printed out or presented as



48

an appropriate graph relating sample mass to temperature. To help

identify the desorption of different species, derivative curves can also be

produced. The results obtained from a sample of Matrex 20 LC silica

gel taken directly from the Perkin Elmer TGA instrument is shown in

figure 33.

Figure 33. Thermogram of Silica Gel

The derivative curve is seen to show three distinctly different desorption

processes. The first takes place from about 30oC to 130oC; the second

between about 200oC and 450oC and the third between about 400oC

and 900oC. The three different desorption processes are distinct and

unambiguous and are similar to those previously identified by Scott and

Traiman (22). The total loss from the sample was about 5%w/w but it

would appear from the TGA curve that the desorption may not be

entirely complete at the temperature of 900oC.

The curve relating mass of water lost (obtained by subtracting each data

point from the initial total mass of sample) against temperature in a

TGA analysis, was a type of desorption isotherm that could be described

by assuming three distinct and separate desorbable species on the silica

surface, each evolving water over a specific temperature range.



49

The first loss of water between 50˚C and 150˚C can be attributed to

physically adsorbed water. The second source that is lost between 150˚C

and 400˚C appears to be strongly held (probably hydrogen bonded)

water. The third and last loss appears to come from the condensation of

the silanol groups to siloxane bonds with the elimination of water. The

amount of water lost during the three desorption processes shown as

desorption curves in figure 34.

Bonded Phases

Bonded phases are formed by reacting the surface hydroxyl groups with

an appropriate reagent to chemical link an organic moiety to the silica

surface. The nature of the organic moiety will determine the type of

interaction that will take place between the solute and the surface. If the

moiety is a hydrocarbon chain, then the interaction will be dispersive, if,

for example, it contains a cyano group, it will be polar and if it carries a

group that can dissociate into ions, the interactions will be ionic.

0 200 400 600 800 1000

0.0

0.2

0.4

0.6

0.8

TE MPERATURE (C)

Wt. Loss 1 Calc

Wt. Loss 2 Calc

Wt. Loss 3 Calc

Figure 34. Graph of Calculated Weight Loss/Temperature

for Each Desorption Stage

The Synthesis of Bonded Phases



50

The particle size will be determined by the nature of the separation and

will vary with the complexity of the mixture and the instrument

characteristics. For example a difficult separation will require high

efficiencies and thus a small particle diameter (e.g. 3 m). If however,

the available column pressure is limited then the particle diameter made

need to be relatively large (e.g. 10 m) to provide adequate flow rate.

The choice of the pore size and surface area of the silica is more

complex. The surface area of silica tends to vary inversely as the pore

size, so the larger the pore size the smaller the surface area. Thus, when

one is defined the other, to some extent, is also specified. The most

efficient bonded phase has the maximum surface coverage. It is

understood, that due to stearic hindrance from the bonded moiety itself,

only a proportion of the silanol groups can be bonded and there is little

that can be done to avoid this problem. However, there are other reasons

for incomplete silanization of the silica. Incomplete silanization can

result from the reagent molecule being excluded from the smaller pores

of the silica. Exclusion can be a particular problem when bonding

relatively large molecular weight materials such as long chain

hydrocarbons onto the silica surface. It is therefore, important to choose

a silica gel that has a relatively large pore size (e.g., a mean pore

diameter of 150Å), which may limit the surface area to between 150 and

250 sq.m per gram and thus, reduce the retentive capacity of the

stationary phase. Nevertheless, this will ensure that the vast majority of

the pores will be accessible to the silanizing reagent. Care must be taken

to ensure that the pore size is not chosen to be so large that the silica gel

is mechanically weak and collapses under pressure during packing or

even during normal use. An alternative approach is to subject the silica

to hydrothermal treatment. Hydrothermal treatment removes silica from

the surface of large pores and deposits it on the smaller pores, which

eventually become blocked and thus, impermeable. This would remove

the small pores, and consequently possible sites for future erosion and

make the large pores even larger and more accessible to the silanizing

reagents.

Bonded Phase Synthesis by Reaction in a Solvent.



51

The solvents normally used in bonded phase synthesis are aromatic

hydrocarbons e.g., toluene that boils at 110˚C or mixed xylenes that boil

138-140˚C. The procedure varies a little depending on the size of the

batch and the type of silanizing reagent. An example of a laboratory

scale synthesis using a chlorosilane is as follows.

10 g of the chosen silica is dried at 250˚C for about 2 hours and

dispersed in a flask containing 100 ml of toluene dried over sodium. If a

mono-chlorosilane reagent is used a small trace of water in the toluene

can be tolerated and can be eventually eliminated by the use of excess

chlorosilane reagent. Under such circumstances the toluene need not be

dried over sodium. If, a dichlorosilane is used however, for example in

the initial step in the synthesis of an oligomeric phase, the presence of

water may cause linear polymerization. Consequently, stringent

precautions must be taken to eliminate all traces of water. A slight

excess of the chlorosilane is then added to the silica dispersion together

with 5 ml of pyridine. The pyridine acts as scavenger for the

hydrochloric acid released during the reaction. The mixture is refluxed

for about 5 hours and the product is then filtered on a sintered glass

filter, washed sequentially with toluene, tetrahydrofuran (THF),

methanol, methanol water (50:50 v/v) and finally with methanol and

dried under suction. The bonded phase now needs end-capping; that is,

any unreacted silanol groups are treated with a small molecular weight

silanizing reagent to react with those hydroxyl groups that were

stearically unavailable to the larger reagent due to exclusion. To end-cap

the product, the bonded phase is refluxed for two hours in a mixture of

100 ml of toluene and 25 ml of hexamethydisilazane. The product is

again filtered free of the reaction liquid mixture and washed sequentially

with, toluene tetrahydrofuran, methanol, methanol water (50:50 v/v) and

finally with methanol and then dried under vacuum. It should be pointed

out that end-capping cannot eliminate the hydroxyl groups that are

stearically hindered by the bonded moiety, or at best, only a small

proportion of them will be removed. It can, however, react with any

readily available hydroxyl groups, particularly those contained in pores

that the original reagent could not enter but to which the smaller capping



52

reagent has access. Capping can also eliminate any hydroxyl groups

attached to the bonded moiety resulting from the presence of dichloro or

tricoloro impurities in the silanizing reagent. After capping, the carbon

content of the product can then be determined to estimate the extent of

reaction.

The carbon content of a bonded phase is often used to determine the

efficacy of bonding, but its value must be used in conjunction with a

knowledge of the surface area of the native silica in order to arrive at a

meaningful conclusion. The amount of material bonded to the silica will

depend, not only on the efficiency of the reaction, but also on the

number of silanol groups that were available with which it could react;

ipso facto it will also depend on the surface area of the parent silica. The

carbon content of the bonded phase is usually determined by micro-

analysis and the result expressed as %w/w of the combined bonded

organic material and the silica gel. Consider a bonded phase where the

carbon content is (y)%w/w coated with a hydrocarbon moiety having (n)

carbon atoms per aliphatic chain (e.g., for the dimethyl octyl brush

phase, n=10). Then, the concentration of aliphatic chains in mols. per

gram of silica (m')will be,

m'

y

100

1

12 n

Consequently, if the surface area of the silica is (A) m2g-1, the number

of mols of aliphatic chains (m) in micromols per square meter will be

given by,

m

y

100

1

12 n

106

A

or, m

833 y

nA (1)

The amount of material bonded per square meter, which indicates the

efficacy of the synthetic process, will be directly proportional to the

carbon content of the product and inversely proportional to the surface

area of the original silica gel.



53

The method of synthesis is very similar for the alkoxysilane reagents.

The same solvents can be used but, as no hydrochloric acid is generated,

there is no need to have pyridine present as a scavenger. The most

reactive alkoxy reagents are the methoxy and ethoxysilanes and their

reaction with a hydroxyl group is accompanied by the release of

methanol or ethanol.

Si CnH2n+ 1(CH3)2

CnH2n+ 1(CH3)2 Si 0

Si O

+HOSi CH3

CH3OH+

The reaction is best carried out in a distillation flask and the methanol

removed as it is formed, the heating rate is adjusted to ensure that the

aromatic solvent is not removed as well. The reaction is allowed to

proceed for about 5 hours and the product then filtered through a

sintered glass filter and washed with the same sequence of solvents as

those used in the chlorosilane synthesis. The product is then refluxed

with the THF/water mixture, filtered and again washed with the

appropriate solvents and dried. The final capping process is also the

same as that employed in the method using the chlorosilanes reagents,

utilizing hexamethyldisilazane as the capping reagent. The alkoxy-

silanes are almost as readily available as the chlorosilanes and are easier

and more pleasant to handle. They are, however, just as hygroscopic as

the chlorosilanes and must be kept under the same anhydrous

conditions. Approximately the same yields can be expected and the

same degree of bonding.

The Fluidized Bed Method for Bonded Phase Synthesis

The concept of synthesizing bonded phases by means of a fluidized bed

reactor was first suggested by Unger (23). The term fluidization



54

describes a contact process in which particulate matter is transformed

into a fluid-like state by a stream of gas or liquid. This state is achieved

by the upward passage of a fluid through the bed of particles to a

velocity at which the drag force acting on the surface of the particles is

equal to the gravitational force downwards. At this point the particles

move apart and become suspended by the fluid flow and the bed is said

to be fluidized.

The fluidized bed apparatus can be operated in such a way that it

allowed far more complex multi-stage syntheses to be carried out rapidly

and with relative ease; e.g., the synthesis of oligomeric phases. The

fluidized bed apparatus is shown in figure 34. The fluidized bed is

enclosed in a tube about 25 cm long, 4.5 cm in diameter situated in a

heating jacket. The temperature of the bed is monitored by three

thermocouples placed in the center and at either end of the bed. A

condenser is situated at the top of the bed, which returns unreacted

silanizing reagent to the vapor generator. The vapor generator consists of

a simple boiling flask that can be provided with a nitrogen stream if the

silanizing reagent does not boil at reasonable temperatures. The

silanizing reagent vapor passes from the vaporizer to a preheater and

then to the base of the fluidizer.



55

Condenser

To Waste

Thermocoup les

Heater Jacket

Silica Gel

Preheater

Sily l Chloride Vaporiz er

St eam Generat er

Valves

Valve

Nit rogen

Flow Meter

Coolant

Coolant Exit

Unreact ed Reagent Return

Figure 34. The Fluidized Bed Apparatus for the Synthesis of

Bonded Phases

A steam generator, similar in form to the silanizing reagent vaporizer, is

also attached to the preheater and consequently, can supply steam to the

fluidized bed if, and when required. This means that the fluidized bed



56

can be used to hydrothermally treat the silica prior to silanization. The

nitrogen stream is continuously monitored by a flow meter. All vapor

and nitrogen streams are adequately fitted with valves to provide the

maximum operating flexibility. The fluidizer is normally loaded with

about 25 g of silica (which provides a fluidized bed about 5 cm long) and

heated to 200oC for about 3 hr. in a fluidizing stream of preheated

nitrogen.

The silanizing reagent is then heated to its boiling point in the vaporizer

or to a temperature where it has a significant vapor pressure and the

reagent blown through the fluidized bed by a stream of nitrogen. Excess,

unreacted reagent is condensed and passed directly back to the

vaporizer. The reaction is allowed to proceed for about 6 hours and then

the reagent vapor flow stopped and replaced by a stream of pure

nitrogen. When reagent is no longer returning to the vaporizer, the bed is

allowed to cool to room temperature with the nitrogen still flowing.

When cool, the nitrogen flow is stopped, and the bonded phase can be

removed from the fluidizer. Alternatively, the silanizing reagent can be

replaced with a capping reagent and the material capped by a similar

procedure and the product then removed.

The fluidized bed synthesis gives a more reproducible product and, at

the same time, eliminates many of the tedious operations that are

involved in the alternative method, such as solvent removal and

recovery, washing procedures and other manipulations. It also offers the

possibility of carrying out very complicated syntheses involving multiple

steps by a relatively simple procedure. However, the technique is more

complicated, requires more elaborate apparatus and needs to be operated

by an experienced technician.

Choosing a Bonded Phase

Different organic moieties can be bonded to all types of silica gel

particles including the very popular spherical variety. One of the most

important features to consider when choosing a bonded phase is the

reproducibility of the product. The packing must obviously achieve the

separation that is required, but unfortunately any bonded phase has a



57

limited lifetime that may range from a few hours, if operated at extremes

of pH, to several months if operated under mild conditions. It follows,

that all columns will eventually need replacing and if a specific

analytical procedure has been established on a particular column, then

the replacement must have as near identical chromatographic properties

as possible. The reproducibility of the column is particularly important

in forensic analysis or for analyses that are carried out to ensure

adherence to regulatory standards as in environmental and pollution

studies. The following tests are recommended as minimal for a new

column and the results should be compared with the data obtained from

the previous column as received.

1/ The column permeability should be measured i.e., the pressure

required to produce a given flow rate e.g., a flow rate of 1 ml per

minute.

2/ The column dead volume should be measured by determining the

retention volume of an unretained solute

3/ The column efficiency should be measured for a set of standard

solutes. If possible, the solutes should be chosen, from those likely to be

present in the samples to be analyzed. Solutes eluting at (k') values of 2,

5 and 10 would be appropriate.

2/ The corrected retention volumes of a series of solutes spanning a (k')

range of 1 to 20 should be determined and their retention ratios

calculated.

All the measurements should fall within 5% of the specified values.

However, these criteria may not be sufficiently stringent for some

forensic purposes and, consequently, the tests given above should be

considered as minimum requirements for litigation purposes..

Types of Bonded Phase

There are three basic types of bonded phases, the brush phases, the bulk

phases and the oligomeric phases. The different phases are produced by

the use of the mono, di- and tri- substituted silanes respectively in the



58

bonding process, e.g. the monochloro, dichloro and tricoloro silanes.

The monochlorosilanes, for instance octyldimethyl-chlorosilane, react

with the hydroxyl groups on the silica surface to produce

dimethyloctylsilyl chains attached to the silica.

CH3CH2CH2CH2CH2CH2CH2CH2 3 2-SiCl + H-O-Si-

CH3CH2CH2CH2CH2CH2CH2CH2 3 2-Si-O-Si-HCl +

)(CH

)(CH

The alkyl chains are thought to stand out from the surface like bristles of

a brush, hence the term brush phase. The extent to which the silanyl

groups are reacted is a subject of debate at this time. It is thought that

the two methyl groups next to the silicon atom hinder the reaction of

adjacent hydroxyl groups with the reagent and thus there will be a

considerable amount of unreacted hydroxyl groups remaining even after

capping. In the extreme, it has been suggested that there is a hydroxyl

group situated between each bonded chain. There is certainly evidence

of some polar interactions with reverse phases, which if completely

covered with hydrocarbon chains should only exhibit dispersive

interactions. However, reverse phases are predominantly dispersive in

character and it would appear that if there are any hydroxyl groups still

present on the surface it is likely that they would be relatively few in

number compared with the bonded moieties.

Oligomeric Phases

The di-substituted silanes such as methyloctyldichlorosilanes can be

used to synthesis the oligomeric bonded phases, which is a far more

complicated procedure (24). The silica is first reacted with the

methyloctyldichlorosilane to link methyloctylchlorosilyl groups to the

surface. The product is then treated with water, which hydrolyses the

methyloctylchlorosilyl groups to methyloctylhydroxysilyl groups with



59

the elimination of hydrochloric acid. The "hydroxy" product is then

reacted with more methyloctyldichlorosilane, attaching another

methyloctylchlorosilyl group to the previous groups.

This process of alternately treating the product with the silane reagent

and then water can be repeated until, in the original synthesis, eight or

ten oligomers are linked to each other and attached to each stearically

available hydroxyl group on the surface. However, it should be noted

that the hydroxylation of the bonded moiety is not the only possible

reaction that can take place with the water. There will be situations

where it will be stearically possible for a water molecule to react with

two adjacent chains and thus produce some cross-linking. However, a

small amount of cross-linking would indeed strengthen the bonded

system and could be advantageous.

The product is finally treated with trimethylchlorosilane or some other

capping reagent to eliminate the last hydroxyl groups formed at the end

of the oligomer. The oligomers are layered over the surface making the

product extremely stable exhibiting almost no polar characteristics

whatsoever. However, due to the complexity of the synthesis [which

needs to be carried out in a fluidized bed for efficient reaction]

oligomeric phases are expensive to manufacture and, consequently, are

not often used and, at this time, are not commercially available.

Nevertheless, this type of procedure, if used to synthesize a reversed

phase, would produce material that would probably be completely

dispersive in character with no measurable polarity.

Bulk Phases

If the silica surface is saturated with water and octyltrichlorosilane is

used as the reagent, reaction occurs with both the hydroxyls of the silica

surface and the adsorbed water causing a cross-linking reaction, and an

octylsilyl polymer can be built up on the surface. The same procedure

can be used as that in the synthesis of oligomeric phases and the

material can be alternatively treated with water and the trichlorosilane

reagent. Layers of bonded phase are built up on the surface but, in this

case, due to the tricoloro function of the reagent, extensive cross-linking



60

occurs. As a result of the polymerization process, the stationary phase

has a chemically cross-linked, multi-layer character and, consequently,

is termed a "bulk" phase. The "bulk" phases are almost as popular as the

"brush" phases as they tend to have a higher carbon content (more

organic material bonded to the surface) and thus provide a little greater

retention and selectivity. "Bulk" phases have about the same stability to

aqueous solvents and pH as the "brush" phases.

Using appropriate organic chlorosilanes, polar or polarizable groups

such as nitriles or aromatic rings can be bonded to the silica to provide

stationary phases covering a wide range of polarities. Bonded ion

exchange materials have also been synthesized, although they are not as

stable to salt solutions and extremes of pH as the ion exchange resins.

Interactions Between 'Brush' and 'Bulk' Reverse Phases and Aqueous

Solvents

The interactions between aqueous solvents and brush reverse phases

differ very significantly from those with a bulk reverse phase at very low

concentrations of solvent.

This difference has been investigated by a number of workers (25-27)

and the basic difference between the two types of phase are shown in

the curves relating retention volume of methanol to the concentration of

methanol in the mobile phase in figure 35. The phases shown are the

RP-18 brush, reverse phase manufactured by E. M. Laboratories, which

had a C18 (dimethyloctadecyl) chain and ODS-3 a bulk reverse phase

which had a C18 (octadecyl) chain and was manufactured by Whatman

Inc. The curves relating retention volume with solvent composition for

the two phases show very different behavior patterns.

The ODS-3 bulk reverse phase behaves in the expected manner, as the

concentration of methanol increases the retention volume of the ethanol

decreases smoothly and continuously up to a concentration of 10%w/v

of methanol. The brush phase, however, behaves in a very unexpected

fashion. The retention volume of ethanol at first increases as the solvent



61

concentration increases, to a maximum at a concentration of about

3%w/v of methanol.

Figure 35. A Graph of the Corrected Retention Volume of Ethanol

against the Concentration of Methanol in the Mobile Phase for a

Bulk and Brush Reverse Phase

At methanol concentrations above 3%w/v the retention values begin to

fall and eventually follow a curve parallel to that for the bulk phase but

somewhat higher. Lochmüller and Wilder (25) carried out some similar

experiments and obtained the same results while Gilpin and Squires (26)

carried out some thermodynamic measurements on the two systems and

showed an anomalous behavior by the brush type phase at very low

concentrations of solvent. It was suggested that the behavior of the brush

phase could be explained as a result of the free chains interacting with

themselves in preference to interacting with the aqueous solvent. In

effect, this was the same phenomena of immiscibility that occurs

between water and a liquid hydrocarbon. The dispersive forces between

the hydrocarbon chains themselves are greater than the forces between

the hydrocarbon chains and the aqueous solvent. As a result the chains



62

interact with themselves and collapse on the surface of the silica. When

the methanol content of the mixture is increased, the solvent becomes

more dispersive in nature and the hydrocarbon chains can then interact

with the solvent and no longer exist in a collapsed state. The situation is

depicted in figure 36.

Brush Phase in Contact wi th Methanol Water Mixture

Brush Reverse Phase in Contact with Pure Water

Figure 36. Effect of Solvent Composition on the Orientation of the

Brush Phase.

The relationship between the retention of ethanol on the brush phase

with water/methanol mixtures, as shown in figure 35, can now be

explained. In pure water the hydrocarbon chains of the brush phase are

collapsed on the surface and thus, the effective surface area of the

stationary phase is much reduced. Consequently, the retention volume of

the solute, being proportional to the available surface area, is also

reduced. As methanol is added to the solvent mixture, the solvent

becomes more dispersive and the hydrocarbon chains can begin to

interact with it and, as a consequence, begin to unfold. The liberation of

the chains from the surface results in an increase in the effective surface

area of the stationary phase and the retention of the solute also starts to

increase. This process continues until there is sufficient methanol in the

solvent for the hydrocarbon chains to totally interact with the solvent

and be completely released from the surface. At this concentration

(about 3%w/v of methanol) the retention volume of the ethanol reaches



63

a maximum. Subsequent increase in methanol concentration merely

increases the interactions of the ethanol with the mobile phase and, by

adsorption of the solvent onto the surface of the reverse phase, reduces

the interactive forces with the reverse phase. Consequently, the retention

volume steadily decreases in the expected manner.

Having explained the behavior of the brush phase it is now interesting to

consider the behavior of the bulk phase. There appears to be no

interaction between the hydrocarbon chains themselves and no change in

surface area at high water concentrations. In fact, the retention of

ethanol falls steadily as the methanol content increases in the expected

manner. The explanation given for this is that the cross-linking that

takes place when the bulk phase is synthesized, keeps the polymeric

chain system rigid and does not allow the individual hydrocarbon chains

to collapse on the surface. As a consequence, the surface area is not

reduced and the retention behavior is normal. Another aspect of the

behavior of the bulk phase at low solvent concentrations is that it does,

in fact, confirm the cross-linked nature of the bulk phase. It would also

appear that for certain solutes, the best reverse phase for operation with

aqueous mixtures containing very little solvent might be a bulk reverse

phase. The retention mechanism on brush type phases under these

conditions might be anomalous.

The Retention Properties of Bulk and Brush Phases.

The equation for the corrected retention volume of a solute (V'r ) (see

Book 6) is as follows,

V'r = KVS or V'r = KAS

where (K) is the distribution coefficient of the solute between the two

phases,

(VS) is the effective volume of stationary phase in the column,

and (AS) is the effective surface area of the packing.



64

The two equations are given to illustrate that the reverse phase system

may be considered as a liquid/liquid or liquid/solid distribution system

where,

VS = dfAS

and (df) is the effective film thickness of the bonded material and

any solvent that may be adsorbed on its surface

It follows that the retention of the solute will depend only on the volume

or surface area of the bonded material. Thus, providing all the bonded

phase is available for solute interaction, the retention volume will be

proportional to the carbon content of the phase. Scott and Kucera (28)

examined a series of commercially available reverse phases and

determined the carbon content of each phase and the retention volume of

a series of solutes on columns packed with each adsorbent. The retentive

properties of the five reverse phase are shown in figure 37 where the

corrected retention volume (V'r) of 2-ethyl anthraquinone is plotted

against carbon content of the reverse phase. It is seen, somewhat

surprisingly, that there is a linear relationship between retention volume

and carbon content of the brush phases (R2, R8, R18). This relationship

can only be expected to occur if all the stationary phase is available to

the solute and the packing procedure is very reproducible so that each

column contains the same amount of packing.

Carbon Content (%w/w)

RP2

RP8

RP18

ODS

ODS2

V'

(ml)

0

5

10

15

20

0 10 20



65

Figure 37. Graph of Retention Volume against Carbon Content

(%w/w)

It should again be stressed that all three reverse phases were produced

from base silicas of very different surface areas and, despite this, the

linear relationship between carbon content and corrected retention

volume remained. This relationship may will depend, not only on the

type of silica gel that is used but also on h bonding process and his

relationship has not been proved generally.

The relative resolution obtained from different reverse phases carrying

diverse carbon contents and extreme chain lengths is shown by the

chromatograms in figure 38. The higher retentive capacity of the bulk

phase ODS2 phase is again clearly demonstrated.

Figure 38. Chromatograms from Brush Phases Carrying Different

Carbon Loads and Different Chain Lengths



66

Short chain reverse phases reduce the extent to which proteins are

denatured in the separation of substances of biological origin, it is seen

by the chromatogram from the C2 reverse phase, that a serious price

must be paid in loss of resolution if the nature of the separation demands

the use of such material. However, the development of the polymer

packings have, at last, partly solved this problem.

In general, because the brush type phases can be synthesized in a more

reproducible manner, particularly if carried out in a fluidized bed, the

brush phases are generally recommended for the majority of

applications. For high retentive capacity and for systems that will be

operated with aqueous solvent mixtures having a very high water

content, the bulk phases might be preferred. The partially reacted, low

carbon content bulk phase may also have special areas of application

particularly in sample preparation.

Macroporous Polymers

Polymeric ion exchange materials were developed for chromatography

in the early sixties resulting in the introduction of macro-porous

polymers (29-31). The advantages of this material lay in the macro-

porous nature of the resin packing, which consisted of resin particles a

few microns in diameter, which, in turn, comprised of a fused mass of

polymer micro-spheres a few Angstroms in diameter. The resin polymer

micro-spheres play the same part as the silica gel primary particles, and

confer on the polymer a relatively high surface area together with a high

porosity. The high surface area provided increased solute retention and

selectivity together with a superior loading capacity and, consequently, a

wide dynamic range of analysis. The material consists of a highly cross-

linked polystyrene resin with about a 50Å pore diameter. In the case of

the ion exchange materials, inorganic groups of appropriate charge were

chemically attached (e.g., by sulfonation). The initial resins

manufactured by Rohm and Haas were called Amberlite but were

largely employed in production processes and, consequently, had very

large particle diameters. Modern macroporous resin-type stationary



67

phases have a range of particle diameters, which can be as small as 2.5

or 3.0 microns. The more popular resin based packings are based on the

co-polymerization of polystyrene and divinylbenzene. The degree of

cross-linking determines the rigidity of the resin and the greater the

cross-linking the harder the resin becomes. Ultimately, at extremely high

cross-linking, the resin becomes brittle. To produce an ion exchange

resin, the surface of the polystyrene-divinylbenzene copolymer is reacted

with suitable reagents and covered with the required ionogenic

interacting groups.

Most cross-linked polystyrene resins employed in LC are the macro-

reticular type and can be produced with almost any desired pore size,

ranging from 20Å to 5,000Å. Underivatized, they exhibit strong

dispersive type interactions with solvents and solutes together with some

induced polarizability arising from the aromatic nuclei in the polymer if

strongly polar solutes are being separated. Consequently, the untreated

resin has found some use as an alternative to the C8 and C18 reverse

phase columns based on silica.



68

Courtesy of TOSOHAAS Inc.

Figure 39. The Separation of a Crude Protein Extract by Exclusion

on a Micro-Reticulated Resin Column

Due to their being stable to extremes of pH, their use for the separation

of peptide and proteins at both high and low pH has been well

established. An example of a macro-reticulated resin phase used as an

exclusion medium in the separation of a crude protein extract is shown

in figure 39. The column, 9TSKgel QC-PAK GFC 399GL0, was 15 cm

long and 8 mm in diameter and the separation was carried out at a flow

rate of 1 ml/min. The mobile phase consisted of 0.5M NaCl in 0.05M

sodium phosphate buffer at pH 7.0. The sample was 5 l of a rat liver

extract. An excellent separation based on molecular size is obtained

from which the molecular weight range of the mixture and even that of

the individual components could be estimated.

LC Mobile Phases

The choice of phase system can be very complex, particularly if

multicomponent mixtures are to be separated. In the first instance the

type of stationary phase needs to be chosen and this choice must be

based on the interactive character of the solutes to be separated. If the

solutes are predominantly dispersive then the stationary phase must also

be dispersive (a reversed phase) to promote dispersive interaction with

the solutes and provide adequate retention and selectivity. If the solutes

are strongly polar then a polarizable stationary phase (one containing

aromatic rings or cyano groups) would be appropriate to separate the

solutes by polar and induced polar interactions. If the solutes are weakly

polar then a strong polar stationary phase would be required (such as

silica gel) to separate the solute by polar interactions.

The mobile phase must be chosen to complement the stationary phase so

that the selected interactions are concentrated in the stationary phase.

Thus, a reversed phase having strong dispersive interactions would be

used with a strongly polar mobile phase (e.g., mixtures of methanol and



69

water acetonitrile and water or tetrahydrofuran and water). In contrast, if

the strongly polar silica gel is selected for the stationary phase then a

strongly dispersive mobile phase would be appropriate (e.g., n-heptane,

n-heptane/methylene chloride or n-heptane with a small quantity of n-

propanol or ethanol). In general the mobile phase must be chosen so that

the selected interactions strongly dominate in the stationary phase and

are minimized in the mobile phase.

Solvent/Solute Interactions with the Silica Gel Surface

In all chromatography systems both the solvent and the solutes interact

with the stationary phase. It follows that, when the silica surface is in

contact with a solvent, the surface is covered with a layer of the solvent

molecules. If the mobile phase consists of a mixture of solvents the

surface is partly covered by one solvent and partly with the other (32).

Thus, any solute interacting with the stationary phase may well be

presented with two, quite different types of surface with which to

interact.

The probability that a solute molecule will interact with one particular

type of surface will be statistically controlled by the proportion of the

total surface area that is covered by that particular solvent.

Dispersive solvents appear to be adsorbed from a solvent mixture on the

surface of silica gel according to the Langmuir adsorption isotherm (33).

Examples of mono-layer adsorption isotherms obtained for benzene,

chloroform and butyl chloride are shown in figure 40.



70

Figure 40. Langmuir Adsorption Isotherms for Benzene, Butyl

Chloride and Chloroform

The adsorption isotherms of the more polar solvents, ethyl acetate,

isopropanol and tetrahydrofuran from n-heptane solutions on to the

silica gel surface did not fit the simple mono-layer adsorption equation

but did fit the bi-layer adsorption isotherm which is a simple extension

of the monolayer formation process. The bi-layer adsorption isotherm

for ethyl acetate on silica gel is shown in figure 41. The curve is

theoretical and the points experimental.

The individual isotherms for the two adsorbed layers of ethyl acetate are

included in figure 41. The two curves, although of the same form, are

quite different in magnitude. The first layer is very strongly held to the

surface and is complete when the concentration of ethyl acetate in the

mobile phase is no more than 1%w/w. As the concentration of ethyl

acetate starts to rise above 1%w/w the second layer is only just being

formed. The formation of the second layer of ethyl acetate is much

slower and obviously the interactions between the solvent molecules

with those already adsorbed on the surface are much weaker than their

interaction with the silica gel silanol groups.



71

0.00

0.05

0.10

0.15

Calculated

Experimental

Firs t Lay er

Sec ond Lay er

Concentration of Ethyl Acetate (% w/v)

Co

nce

ntr

ati

on

on

Sil

ica

(g/g

)

0 1 2 3 4

First Layer

Se cond Laye r

Combined Layer

Figure 41. The Individual and Combined Adsorption Isotherms for

Ethyl Acetate on Silica Gel

If it is assumed that the total area covered by the first layer of ethyl

acetate will be very similar to the area covered by the second layer, then

only about one third of the second layer is complete at an ethyl acetate

concentration of about 4%w/v. In contrast, the first layer is virtually

complete at an ethyl acetate concentration of 1%w/v. It should be

pointed out that solvent layer formation on the surface of the silica is not

necessarily restricted to binary systems and multi-layers are quite

feasible.

Solute Stationary Phase Interactions

There are basically two types of interaction that can take place between

a solute and the silica gel surface. Firstly, the solute molecule can

interact with the adsorbed solvent layer and rest on the top of it. This

type of interaction is called sorption interaction and occurs when the

molecular forces between the solute and the silica are relatively weak

compared with the forces between the solvent molecules and the silica.

The second type of interaction is where the solute molecules displace the



72

solvent molecules from the surface and interact directly with the silica

gel itself, for example, the silanol groups.

SORPTION

SO LUTE

CHLOROFORM n-HEPTANE

DISPLCACEMENT

CHLOROFORM n-HEPTANESOLUTE

SOLUTE

Figure 42. Diagram Depicting Sorption and Displacement

Occurring on the Silica Surface.

These two types of interaction are shown in figure 42. Displacement

would occur if the solute was strongly polar such as an alcohol, which

would interact more strongly with the polar silanol group than the

dispersive chloroform layer. Sorption is depicted as a solute molecule on

interacting with each solvent layer and can not interact strongly enough

with the silica gel surface to displace the solvent..

Mobile phases consisting of mixtures of polar and dispersive solvents

frequently produce surface bi-layers when used with silica gel as a

stationary phase and therefore a far more complicated set of interactive

possibilities exist. These possibilities are depicted in figure 43.



73

So lute i ntera cting

with second l ayer of

sol vent (B) SORPT ION


with first laye r of

sol vent (B) by

displaci ng so lvent

fro m second layer

DISPLACEM ENT

Displace d

So lvent (B)

Displace d

So lvent (B)Displace d

So lvent (A)


directly with f irst la yer

of solve nt (B) SORPTION


with sil ica surface by

displaci ng so lvent (A)

fro m first laye r

DISPLACEM ENT


with sil ica surface by

displaci ng so lvent (B)

fro m first laye r

DISPLACEM ENT So lute i ntera cting

directly with l ayer of

sol vent (A) SORPT ION

Figure 43. Different Types of Solute Interaction that can occur on

Silica Surfaces Covered with a Solvent Bi-layer

The surface offers the opportunity of a number of sorption and

displacement processes that can take place between the solute and the

stationary phase surface. There are three different surfaces on which a

molecule can interact by sorption and three different surfaces from

which molecules of solvent can be displaced and allow the solute

molecule to penetrate. In any separation all the alternatives are possible

but it is more likely that for one particular solute, one type of interaction

will dominate. the various types of interaction are depicted in figure 43.

Where there are multi-layers of solvent, the solvent that interacts directly

with the silica surface is the most polar, and consequently constitutes the

first layer. Depending on the concentration of the polar solvent the next

layer may be a second layer of the same polar solvent as in the case of

ethyl acetate. If, however, the quantity of polar solvent is limited, then



74

the second layer might consist of a less polar component of the solvent

mixture. If a ternary mixture of solvents is used, the nature of the

surface, and the solute interactions with the surface can become very

complex indeed. In general the stronger the polarity of the solute the

more likely it is to interact with the surface by displacement even to the

extent of displacing both layers of solvent (one of the alternative

processes that is not depicted).

Solvent/Solute Interactions with the Reversed Phase Surface

Solvents interact with the surface of a reverse phase in a similar manner

to the surface of silica gel. In figure 44, the adsorption isotherms for a

series of aliphatic alcohols are shown. The effect of the carbon chain-

length of the alcohol on the strength of the adsorption is clearly seen

from the shape of the curves. The most strongly adsorbed alcohol,

butanol, (the alcohol with the longest chain and, thus, the most

dispersive) has only a four carbon chain and yet the surface is

completely covered when the butanol concentration is only about

2%w/v. Thus, any component of the mobile phase with a hydrocarbon

chain length of four or more, will be rapidly adsorbed and modify the

reverse phase surface extensively and, consequently, the magnitude of

solute retention.

The curves shown in figure 44 only cover a range of 0 to 0.05 g.ml-1. In

order to show the shapes of the adsorption isotherms for the higher

alcohols in proportion to those of the lower alcohols with reasonable

clarity, the same curves are shown in figure 45 for an alcohol

concentration range of 0-100% (which is approximately 0-0.8g/ml).



75

0 0.01 0.02 0.03 0.04 0.05

0

Concentrati on of Alcohol (g/ml)

Methanol

Ethanol

n-Propanoln-Butanol

Ma

ss o

f So

lvent

on

Surf

ace

(g

/cm

)

2

It should be noted that the mass adsorbed is expressed as g.cm-2

Figure 44. The Adsorption Isotherms of a Homologous Series of

Aliphatic Alcohols over the Concentration Range of 0 to 0.0.5 g.ml-

1

The weak nature of the methanol adsorption, relative to the other

alcohols is clear and it is seen that the surface of the reverse phase is

being modified over one third of the methanol concentration range. The

reverse phase surface can be modified in a controlled manner, over the

range of 0 to about 40 % methanol, but between methanol

concentrations of 40% and 100 % the nature of the reverse phase

surface remains sensibly constant and it is the solute interactions in the

mobile phase that are progressively modified. Acetonitrile and

tetrahydrofuran behave in a similar manner but their adsorption

isotherms are closer in magnitude to those of ethanol than of methanol.

The types of interactions that can take place between the solute and the

reverse phase are similar to those that can take place between the solute

and the silica gel surface. Solutes can interact by the sorption process,

the displacement process or a combination of both. The same rules

apply; if the solvent interacts more strongly with the surface than the



76

solute then the solute interacts with the adsorbed layer of solvent by

sorption.

3

4

0 0.16 0.32 0.48 0.800.64

Concentration of Alcohol (g/ml)

0

1 Butanol 2 Propanol

3 Ethanol 4 M ethanol

1

2

Mass

of

Solv

ent

on

Su

rface

(cm

)

-2

Figure 45. The Adsorption Isotherms of a Homologous Series of

Aliphatic Alcohols

If, on the other hand, the solute interacts more strongly with the reverse

phase than the layer of solvent molecules then the solute will displace

the solvent and interact directly with the surface by displacement. In,

general, those solutes that elute early in the chromatogram will interact

by sorption, those that elute late in the chromatogram will interact by

displacement and at some intermediate point in the elution scale, solute

stationary phase interactions will probably involve both sorption and

displacement. Bi-layer adsorption is also possible with reverse phases

but, at this time, experimental evidence of this does not appear to be

available in the literature.

Molecular Interactions in the Mobile Phase

The 'elutive' capacity of the mobile phase (as opposed to the 'retentive'

capacity of the stationary phase) depends on the strength of the different

interactions that can take place in the mobile phase and the probability

of a particular interaction occurring. Purnel and Laub (34)



77

experimentally demonstrated in GC, that for a stationary phase

consisting of a pair of non-associating liquids, the distribution

coefficient of a solute was linearly related to the volume fraction of

either liquid. This relationship indicated that the volume fraction of a

solvent in a liquid mixture determined the probability of interaction.

Much the same as the partial pressure of gas determines the probability

of collision. This relationship was challenged by a number of workers

who also demonstrated that for certain solvent pairs this linear

relationship broke down. However, it was also shown (35,36) that the

nonlinear relationship occurred when there was strong association

between the components of the mixture which resulted in a ternary

mixture containing the two components and the associate of the two

components the concentration of which depended on the equilibrium

constant and the experimental conditions.

Thus, for a mobile phase consisting of a binary mixture of solvents, as

the retention volume will be inversely proportion to the elutive capacity

of the mobile phase it will be also inversely proportional to the volume

fraction of either component providing there is no strong association

between the components. This was experimentally demonstrated by Katz

et al. (35) who employed a liquid/liquid distribution system using water

and a series of immiscible solvent mixtures as the two phases and

measured the absolute distribution coefficient of a solute for different

mixtures.

The solute they used was n-pentanol and the immiscible solvent

consisted of mixtures of n-heptane and chloroheptane, n-heptane and

toluene and n-heptane and heptyl acetate. The two-phase system was

thermostatted at 25oC and, after equilibrium had been established, the

concentration of solute in the two phases was determined by GC

analysis. The results they obtained are shown in figure 46. It is seen that

linear relationships between solvent composition and distribution

coefficient was obtained for all three solvent mixtures simulating the

results that Purnell and Laub obtained in their gas chromatography

experiments.



78

Figure 46. Graph of Distribution Coefficient against Solvent

Composition.

Aqueous Solvent Mixtures

When the relationship between the distribution coefficient of a solute

and solvent composition, or the corrected retention volume and the

solvent composition, was tested with aqueous solvent mixtures it was

found that the relationship identified by Purnell and Laub and Katz et al

failed. It was suspected that the failure was due to the solvent strongly

associating with the water and, in fact, an aqueous solution of methanol,

for example, contained methanol, water and methanol associated with

water. The solvent mixture was thus, a ternary system and thus the

linear relation ship between the volume fraction (before mixing) and

retention would not be expected to hold.

The association of methanol and water was examined by Katz,

Lochmüller and Scott (36) using volume change on mixing and

refractive index data and established that the methanol/water solvent



79

system was indeed a complex ternary system. They calculated both the

association equilibrium constant and the distribution of the different

components of a methanol water mixture form zero to 100% methanol.

The curves they obtained are shown in figure 47.

Figure 47. Diagram of the Ternary Solvent System for

Methanol/Water Mixtures

It is seen from figure 47 that there are three distinct ranges of methanol

concentration where the solvent will behave very differently. From zero

to 40%v/v of methanol in the original mixture, the solvent will largely

behave as though it were a binary mixture of water and methanol

associated with water. From 40%v/v to 80%v/v of methanol in the

original mixture, the solvent will predominantly behave as though it

were a ternary mixture of water, methanol and water associated with

methanol. From 80%v/v to 100%v/v of methanol in the original mixture,



80

the solvent will again behave as though it were again a binary mixture

but this time a mixture of methanol and water associated with methanol.

The curves shown in figure 47 explain some of the unique

characteristics of mobile phases consisting of methanol water mixtures

when used in reversed phase LC. From figure 47 it is seen that when the

original mixture contains 50%v/v of methanol there is little free

methanol available in the mobile phase to elute the solutes as it is mostly

associated with water. Subsequently, however, the amount of methanol

unassociated with water increases rapidly in the solvent mixture and this

rapid increase must be accommodated by the use of a convex gradient

profile when employing gradient elution. The convex gradient will

compensate for the strongly concave form of the unassociated methanol

concentration profile shown in figure 47 which will be the strongest

eluting component of the mobile phase. The strong association of

methanol with water could also account for the fact that proteins can

tolerate a significant amount of methanol in the mobile phase before

they become denatured. It is clear that this is because there is virtually

no unassociated methanol present in the mixture, which could cause

protein denaturation since all the methanol is in a deactivated state by

association with water.

Katz, Lochmüller and Scott also examined acetonitrile/water, and

tetrahydrofuran(THF)/water mixtures in the same way and showed that

there was significant association between the water and both solvents

but not to the same extent as methanol/water. At the point of maximum

association for methanol, the solvent mixture contained nearly sixty

percent of the methanol/water associate. In contrast the maximum

amount of THF associate that was formed amounted to only about 17%

and for acetonitrile the maximum amount of associate that was formed

was as little as 8%. It follows that acetonitrile water mixtures would be

expected to behave more nearly as binary mixtures than methanol/water

or THF/water mixtures.

Katz et al. measured the distribution coefficient of benzene between n-

hexadecane and some methanol/water mixtures. Using the data from



81

figure 47, they plotted the distribution coefficient of benzene against the

volume fraction of methanol unassociated with water. The results they

obtained are shown in figure 48.

Figure 48. Graph of Distribution Coefficient of Benzene Between

Methanol and Water Mixtures and n-Hexadecane against Volume

Fraction of Methanol Unassociated with Water in the Aqueous

Mixture

From the results in figure 48, benzene does not appear to interact with

methanol associated with water or water itself but solely with methanol.

The linear curve is obtained with zero intercept confirming the validity

of this dependence. In fact, the methanol associated with water plays no

significant part in competing for the benzene against the dispersive

interactions of the n-hexadecane.

Chiral Stationary Phases



82

There are basically five general types of chiral stationary phase in

common use in LC. The first is the protein based stationary phase.

These stationary phases usually take the form of natural proteins bonded

to a silica matrix. As they are proteins, they contain a large number of

chiral centers and are known to interact strongly with small analytes

exhibiting strong chiral selectivity. There are specific interactive sites

that provide chiral selectivity, but there are many more sites that only

contribute to general retention. These other sites can be deactivated with

mobile phase additives (e.g. octylamine), which reduces the overall

retention and increases the chiral selectivity.

The second type consists of relatively small molecular weight chiral

substances bonded to silica Pirkle (37). Each bonded group has a limited

number of chiral centers available but, due to their small size, there can

be a large number of groups bonded to the silica (as opposed to much

larger complex chiral moieties). It follows, that a relatively high

probability is maintained of the solute interacting with a chiral center.

The advantage of the Pirkle chiral phases is that, as the overall

interacting molecule is small, the solutes are not strongly retained and

thus the chiral selectivity becomes the dominant factor. The third type is

based on polymers of cellulose and amylose, which were developed by

Okamato (38). These are derivatized to link appropriate interactive

groups to the cellulose polymer, which is then coated onto a silica

support. The fourth type is based on the macrocyclic glycopeptides

introduced by Armstrong (39). These are materials that also contain a

large number of chiral centers, together with molecular cavities in which

solute molecules can enter and interact with neighboring groups.

The spatial character of the solute will determine the degree of entry and

consequently the proximity of interaction, which, in turn, will determine

the energy of interaction and the magnitude of the retention. Finally, the

fifth group contains the cyclodextrin-based materials that control

retention in a similar manner to that previously described for GC. In LC,

the cyclodextrin stationary phases are bonded to a support such as silica

and are prepared using similar techniques to those for making reverse

phases. The more recent and most effective stationary phases are



83

without doubt those based on the macrocyclic glycopeptides and the

cyclodextrins.

Macrocyclic Glycopeptide Phases

The concept of using macrocyclic glycopeptides as chiral stationary

phases was first introduced by Armstrong [39]. One method of

preparation is to covalently bond Vancomycin to the surface of silica gel

particles. Vancomycin contains 18 chiral centers surrounding three

'pockets' or 'cavities' which are bridged by five aromatic rings. Strong

polar groups are proximate to the ring structures to offer strong polar

interactions with the solutes. This type of stationary phase is stable in

mobile phases containing 0–100% organic solvent. The proposed

structure of Vancomycin is shown in figure 49.

O

CH3

NH2

CH3

OH

OO

NN

N

NN

N

O

OH

H

HOH

HOOC

HH

OO

O

OO

H

H H2NC

O

Cl

NHCH3

H H

O

O

O

OHHO

HOCH2

OHOH HO

H

Cl

A

BC

A, B and C are inclusion cavities. Molecular weight 1449. Chiral centers 18. pK's

2.9, 7.2, 8.6, 9.6, 10.4, 11.7. Isoelectric point 7.2

Courtesy of ASTEC Inc.

Figure 49 The Proposed Structure of Vancomycin



84

Vancomycin is very stable with a relatively high sample capacity, and,

when covalently bonded silica gel has multiple linkages to the silica gel

surface. It can be used as a reversed phase, with mobile phases having a

high water content, or, alternatively, as polar stationary phase with a

mobile phase of high solvent content (e.g., when used as a reversed

phase, strongly polar THF–water mixtures are very effective mobile

phases). Conversely, when used as a polar stationary phase, n-hexane–

ethanol mixtures are appropriate. Vancomycin has a number of ionizing

groups and thus can be used over a range of different pH values (pH 4.0

to 7.0) and exhibit a wide range of retention characteristics and chiral

selectivities. Ammonium nitrate, triethylammonium acetate and sodium

citrate buffers have all been used satisfactorily with this stationary

phase. An example of the use of the stationary phase to separate the

enantiomers of 3-methyl-5-phenylhydantoin is shown in figure 50.

to 1.85 min. to 2.80 min.

k1 0.78 k1 1.57

k2 1.32 k2 2.13

1.69 1.35 R 2.18 R 3.0



85


Figure 50. The Separation of the Enantiomers of 3-Methyl-5-

Phenylhydantoin Using Polar and Dispersive Interactions

The separation was carried out under two conditions, the first used pure

ethanol as the mobile phase, which is strongly dispersive, and in the

second, the mobile phase that contains 90% of water. In the first case,

the ethanol provides strong dispersive interactions in the mobile phase,

which would significantly exceed any dispersive interactions involved

with the stationary phase. Consequently, the remaining dominant

retentive forces will be polar or ionic. In the second case, the mobile

phase is predominantly water and thus provides strong polar interactions

with the solute but weak dispersive interactions. It also follows, that the

dispersive forces will dominate in the stationary phase. These two

examples demonstrate the useful flexibility of Vancomycin.

Another macrolytic glycopeptide used in chiral chromatography is the

amphoteric glycopeptide Teicoplanin which is commercially available

under the trade name of CHIROBIOTIC T.



86

HO

O

NH2

O

HO

Cl

HHO

H

O

H

O

N NNH

O

H

B

A

OCl

N

HO

HHO

C

OHOHO

NHH

NH

O

HOOC

HD

HO

HO

HO NHR

CH2OH O

CH2OH O

HO

HOHNCOCH3

OH

CH2OHOH

OH

O

A, B, C and D are inclusion cavities. Molecular weight 1885. Chiral centers 20,

Sugar moieties 3, and R is CH3-decanoic acid


Figure 51. The Proposed Structure of Teicoplanin

This material can also be bonded to 5 m silica gel particles by multiple

covalent linkages. Teicoplanin contains 20 chiral centers surrounding

four molecular 'pockets' or 'cavities'. Neighboring groups are strongly

polar and aromatic rings provide ready polarizability. The proposed

structure of Teicoplanin is shown in figure 51. This stationary phase is

claimed to be complementary to the Vancomycin phase and can be used

with the same types of mobile phase, one often providing chiral

selectivity, when the other does not. Teicoplanin can be used in a

reversed phase mode using strongly polar mixtures such as

acetonitrile/aqueous buffer : 10/90 v/v, THF/aqueous buffer 10/90 : v/v,

and ethanol/aqueous buffer : 20/80 v/v). It can also be used as a polar

stationary phase using n-hexane/ethanol mixtures as the mobile phase. In

some cases it is advisable to control the pH even when the solutes are

not ionic, suitable buffers being ammonium nitrate and triethylamine



87

acetate. The separation of the Propranolol enantiomers on Teicoplanin is

shown in figure 52.

Figure 52 The Separation of the Enantiomers of Propranolol

Employing Different Acid/Base Ratios

The column used was 25 cm long, 4.6 mm I.D., packed with Chirobiotic

T. The mobile phase was methanol containing acetic acid and

triethylamine in the concentrations shown in figure 52. The column was

operated at room temperature and at a flow rate of 2 ml/min.

Teicoplanin is stable over a pH range of 3.8 to 6.5 although it can be

used for limited periods of time outside this range. By suitable choice of

mobile phase, Teicoplanin can be used in the reversed phase mode.

Another macrolytic glycopeptide that has been used as a chiral

stationary phase, is the glycopeptide, Avoparcin [40]. This stationary

phase is available as CHIROBIOTIC A. Avoparcin is an antibiotic

complex produced by Streptomysces candidus. In particular, it is used to

prevent necrotic enteritus in chickens. The structure of avoparcin is

shown in figure 53.



88

Figure 53. The Structure of Avoparcin

There are two forms of Avoparcin the unsubstituted -Avoparcin

structure and the chlorinated structure -Avoparcin, the molecular

weights being 1909 and 1944 respectively. The ratio of -Avoparcin to

-Avoparcins is about 1:4. The aglycon portion of Avoparcin contains

three connected semie rigid macrocyclic rings (one 12-membered, and

two 16-membered), which form a pocket providing possible solute

inclusion. The glycopeptide contains seven aromatic rings with four

phenol moieties, four carbohydrate chains, 16 hydroxyl groups, one

carboxylic acid, two primary amines, one secondary amine, six amide

linkages, two chlorine atoms for -Avoparcin (only one for -

Avoparcin) and 32 stereogenic centers.

It is clear that, there is a wide diversity of interactive possibilities

ranging from weak and strong dispersive interactions, to polar

interactions that span from induced dipole interaction, through dipole–



89

dipole interaction, to strong hydrogen bonding. In addition, at the right

pK, basic and acidic ionic interactions can also be invoked. More

importantly, with 32 stereogenic centers the probability of interaction

between chiral centers of solute and stationary phase is relatively high.

Cyclodextrin

The cyclodextrin based chiral stationary phases are some of the more

popular materials used for contemporary chiral separations. One of their

advantages lies in their use with all types of solvent. They can be used

very effectively in the reversed phase mode and, as well as being usable

as a normal phase. The cyclodextrins and their derivatives have been

widely used for all types of chiral separations and can often be used for

preparative separations. Cyclodextrin-based phases are readily available,

covalently bonded to spherical silica gel particles 5 m in diameter. The

cyclodextrins are produced by the partial degradation of starch followed

by the enzymatic coupling of the glucose units into crystalline,

homogeneous toroidal structures of different molecular size. The

molecular structure of , , and cyclodextrins are shown in figure 54.

The alpha-, beta- and gamma-cyclodextrins and have been shown to

contain 6 (cyclohexamylose), 7 (cycloheptamylose) and 8

(cyclooctamylose) glucose units, respectively. These cyclic, chiral, torus

shaped macromolecules contain the D(+)-glucose residues bonded

through -(1-4)glycosidic linkages.



90


Figure 54 The Molecular Structure of , , and Cyclodextrins

The mouth of the torus-shaped cyclodextrin molecule has a larger

circumference than at the base and is linked to secondary hydroxyl

groups of the C2 and C3 atoms of each glucose unit (see figure 55). The

primary hydroxyl groups are located at the base of the torus on the C6

atoms. As these hydroxyl groups are free to rotate, they partially block

the base aperture. The size of the cavity increases with increasing

number of glucose units (figure 54). The secondary hydroxyl groups can

be reacted with appropriate reagents to introduce further interactive

character to the cyclodextrin molecule. The very effective chiral

characteristics of the cyclodextrins structures arise from the many chiral



91

centers they contain, for example, -cyclodextrin has 35 stereogenic

centers.

Courtesy of Supelco

Figure 55 A Molecular Model of Cyclodextrin

When the , or cyclodextrins are derivatized, the hydroxyl group on

the 2-position reacts first. However, the derivative is still size selective

and interaction will be determined by the size and functional groups

contained by the interacting molecule. Derivatizing the 6-hydroxyl

position has little or no effect on chiral selectivity but does enhance the

loading capacity of the stationary phase. This position is used for

anchoring the cyclodextrins to silica gel in the preparation of LC

stationary phases.



92

The cyclodextrins have found a wide field of application in chiral

chromatography and there arc many applications in the literature The

following is a simple example of the use of this stationary phase.

In order to carry out in vivo pharmacological profiling of enantiomeric

drugs, the direct analysis of biological fluids is required to reduce

sample preparation time, and the chance of enantiomeric change.

Unfortunately, many chiral stationary phases, such as the Pirkle types

phases and the derivatized cellulose phases, demand the use of mobile

phases that are incompatible with the biological fluids. Haginaka and

Wakai [41] suggested that the silica should first be reacted with (3-

glycidoxypropyl)trimethoxysilane, to cover a significant part of the silica

surface with 'spacers', and then the remaining silanol groups reacted

with cyclodextrin–carbamoylated triethoxysilane to attach the chiral

agent.

Courtesy of J. Liq. Chromatogr. [Ref.17]

Figure 56. The Separation of the Hexobarbital Enantiomers

Contained in Blood Serum by Direct Injection on a Cyclobond

Phase



93

Subsequent treatment, would convert the epoxy group to diols that

would insulate the analytes from the silica surface. Stalcup and Williams

[42] analyzed a series of these type of materials and found that there was

about 10 times as much spacer on the surface (ca. 2 mol/m2) as there

was derivatized cyclodextrin (ca. 0.2 mol/m2). This type of chiral

stationary proved to be very effective. The separation of the enantiomers

of hexabarbital on this stationary phase by the direct injection of blood

serum is shown in figure 56. Chromatogram A was obtained after 20

injections of serum and chromatogram B after 60 consecutive injections

of blood serum. It is seen that here is very little column deterioration and

that, although the tail of the major peak has become a little extended

after 60 injections, the column could still be used very effectively for the

analysis.

Liquid Chromatography Applications

Liquid chromatography has been used in an extremely wide range of

analyses and it is impossible to give a comprehensive set of examples

that would illustrate its wide applicability. The following are a few LC

analyses that may indicate the scope of the technique and give the reader

some idea of its importance and versatility.

An example of the use of reversed phase chromatography (employing a

C8 column) for the separation of some benzodiazepines is shown in

figure 57. The column used was 25 cm long, 4.6 mm in diameter

packed with silica based, C8 reverse phase packing particle size 5 .

The mobile phase consisted of 26.5% v/v of methanol, 16.5%v/v

acetonitrile and 57.05v/v of 0.1M ammonium acetate adjusted to a pH

of 6.0 with glacial acetic acid and the flow-rate was 2 ml/min.

The column efficiency available at the optimum velocity would be about

15,000 theoretical plates. The retention time of the last peak is about 12

minutes (i.e., a retention volume of 24 ml). At a flow rate of 2 ml/min.,

the mobile phase velocity will be well above the optimum and so the

maximum efficiency has not been realized. The general technique used

when there are more theoretical plates available than required is to



94

increase the flow rate until the separation required is just realized. This

procedure trades efficiency for time and allows the separation to be

achieved in the minimum time given the column and phase system that

has been chosen.

Courtesy of Supelco Inc

.

Figure 57. The Separation of Eight Diazepines Employing a C8

Reverse Phase Column

The mobile phase is buffered appropriately to complement the

dissociation constants of the solutes. A mixture of methanol and

acetonitrile is employed, the acetonitrile being used to increase the

dispersive interactions in the mobile phase. The reason for the particular

solvent mixture is not clear and it would appear that the separation

might be achieved equally well by using a stronger solution of methanol

alone or a more dilute solution acetonitrile alone. There is no particular



95

advantage to one solvent mixture over another except for the fact that

'waste' acetonitrile produces greater solvent disposal problems than

methanol. Another example of the use of reversed phase

chromatography for the separation of mixture of growth regulators (a

C18 reverse phase column) is shown in figure 58.

Courtesy of Supelco Inc.

Figure 58. The Separation of a Mixture of Growth Regulators on a

C18 Reverse Phase C18 Column

The packing is silica based but is contained in a short column 3.3 cm

long, 4.6 mm in diameter and packed with particles 3 m in diameter.

The expected efficiency of the column (when operated at the optimum

velocity) would be about 5,500 theoretical plates. This is not a



96

particularly high efficiency and so the separation depends on the phases

chosen to provide the necessary selectivity and an appropriate gradient

program.

The selectivity was achieved using a complex mixture of ionic and

dispersive interactions between the solutes and the stationary phase and

ionic, polar and dispersive forces between the solutes and the mobile

phase. The initial solvent in the gradient program was a 1% acetic acid

and 1 mM tetrabutyl ammonium phosphate buffered to a pH of 2.8. The

tetrabutyl ammonium salt would be adsorbed strongly on the reverse

phase and thus acted as an adsorbed ion exchanger. During the program,

acetonitrile was added to the solvent and initially this increased the

dispersive interactions between the solute and the mobile phase.

As the acetonitrile concentration became higher, however, the tetrabutyl

ammonium salt would be desorbed from the reverse phase reducing the

ionic interactions of the solutes with the stationary phase. At even higher

concentrations of acetonitrile, the tetrabutyl ammonium salt would be

completely desorbed and the interactions of the solutes with the

stationary phase would become almost exclusively dispersive. This is an

example where a complex phase system was necessary because there

was limited column efficiency available. It is likely that a column with

intrinsically more efficiency might achieve the separation with a much

simpler solvent system and a more straightforward solvent program.

An example of the use of native silica is given by for the analysis of

Darvocet® and its generic equivalent formulation. Darvocet® is an

acetaminophen product in which the active ingredient (and other

material in the medicine) are weakly polar and, consequently, lend

themselves to separation on a strongly polar stationary phase such as

silica gel. The analysis is depicted in figure 59. The analysis is

completed in less than 4 minutes using a short column 3.3 cm long and

4.6 mm in diameter. The silica packing had a particle size of 3

providing a maximum efficiency of about 5,500 theoretical plates.



97

In order to identify the impurities, the column had to be significantly

overloaded. Despite this, the impurities were well separated from the

main component and a substance was shown to be present in the generic

formulation that was not in the Darvocet®. The mobile phase was

98.5% dichloromethane with 1.5% v/v of methanol containing 3.3%

ammonium hydroxide. The ammoniacal methanol deactivated the silica

gel but the interaction of the solutes with the stationary phase would still

be polar in nature. In contrast solute interactions with the methylene

dichloride would be exclusively dispersive.


Figure 59. The Analysis of Acetaminophen Formulations

Ion chromatography can be used in a number of novel ways and

employing the appropriate conditions can even be used to separate

mixtures where the components are not ionic or do not normally

produce interactive ions in aqueous solution. An example of this type of

separation is the analysis of saccharide mixtures using ion exchange



98

interactions. An illustration of such a separation is given in figure 60.

The saccharides are reacted with a borate with which saccharides

readily forms complex anions. The procedure for making the complex is

simple and is achieve by merely including a borate buffer in the mobile

phase. The process is, in fact, a form of 'in-line' derivatization.

Courtesy of TOYO SODA Manufacturing Co., Ltd.

Figure 60. The Separation of a Saccharide Mixture by Ion

Exchange Chromatography

The column packing was a strong anion exchange resin designated as

TSKgel Sugar AXG. It had a particle diameter of 10 and contained

quaternary ammonium ions as the ion exchange moiety. The column



99

was 15 cm long, 4.6 mm in diameter and had a potential efficiency of

about 7,500 theoretical plates. The mobile phase consisted of three

borate buffer solutions, which were used in a stepwise gradient. The first

buffer solution was a 0.5 M borate buffer (pH 7.7), the second a 0.7 M

borate buffer (pH 7.3) and the last a 0.7 M borate buffer (pH 8.7) and

the flow rate was 4 ml/min.

The last example clearly broaches a decidedly different approach to the

sample separation. It is just as feasible for the solutes to be modified to

suit a particular phase system as it is to choose or modify a phase system

to suit the solutes. This emphasizes the wide range of variables and

alternative approaches that liquid chromatography provides for the

analyst.

Before ending an example needs to be included that utilizes Micro-

reticulated polystyrene gels as a stationary phase. Micro-reticulated

polystyrene gels, as already discussed, are formed from cross-linked

styrene-divinyl benzene polymers and can be manufactured with a wide

range of different pore sizes. Micro-reticulated polystyrene gels,

however, are not used solely for the separation of large multi-interactive

molecules (e.g., protein and polypeptide mixtures), but can be used for

the same type of analyses as silica gel.



100


Figure 61. Separation of Some Phthalate Esters by Exclusion

Chromatography on Styrene-Divinyl Benzene Based Gel

Due to their method of manufacture, the pore volume of the micro-

reticulated polystyrene gels may be significantly lower than that of silica

gel. Consequently, some gels may have neither the peak capacity nor the

loading capacity normally experienced with silica. However, this is

generally not a problem in analytical LC. An example of the separation

of some phthalate esters is shown in figure 61. The column was 30 cm

long and 7.8 mm diameter. In exclusion chromatography, a large

diameter column is necessary to provide adequate peak capacity (see

Book 6). The particle diameter of the packing was 6 m and thus, at the

optimum velocity, an efficiency of about 25000 theoretical plates should

be produced. It is seen in figure 61 that the dead volume time is about

19 minutes and a flow-rate of 1.0 ml/min this would be equivalent to a

dead volume of 19.5 ml.

Courtesy of Asahipak Inc.



101

Figure 62. The Separation of a Standard Protein Mixture by

Exclusion Chromatography on a Vinyl Alcohol-Styrene Co-

Polymer Hard Gel

A more typical application of micro-reticulate resins in exclusion

chromatography is shown by the separation of a standard protein

mixture depicted in figure 62.

The column was 50 cm long and 7.6 mm wide giving a total column

volume of about 22.7 ml. This conflicts with the value for the dead

volume, which appears to be about 22 ml (the retention time of 22 min.

at a flow rate of 1 ml/min.). Consequently, it would appear that the resin

occupied no volume in the column. This paradox could be explained on

the basis that the separation was not achieved solely by exclusion. The

dispersive and polarizable nature of the resin was contributing dispersive

and possibly slight polar interactions with the solutes increasing their

retention beyond that expected from exclusion alone. This would be

supported by the fact that the mobile phase contained no solvent and

consisted of a buffer solution containing 0.1 M sodium phosphate and

0.3 M sodium chloride (pH 7.0), which would have minimum dispersive

interactions with the solute. In contrast, the hydrocarbon matrix of the

styrene gel would be capable of exerting strong dispersive interactions

with the dispersive groups present in the proteins. The use of mixed

retention systems to achieve the separation is perfectly acceptable and

occurs in most distribution systems. The stationary phase was GFA-50,

the number 50 referring to the columns length. The pore size was

defined in terms of the minimum molecular weight of a totally excluded

solute which was given as 300,000 and the material was specifically

prepared for exclusion chromatography.

The analyst should be aware that generally, when working with large

molecules (particularly proteins), column efficiencies may, under some

circumstances, be as little as 10% of the expected value and this must

be taken into account when choosing the column and the phase system.

References



102

1. M.S.Tswett, Tr .Protok . Varshav . Obshch . Estestvoispyt Otd .

Biol . 14(1905).

2. R. Willstatter and A. Stoll, Utersuchungenuber Chlorophy ,

Springer, Berlin 1913

3. A. J. P. Martin and R. L. M. Synge , Biochem . J . 35(1941)1358.

4. A. Tiselius and D. Claesson, Ark. Kemi. Mineral. Geol.

15B(No 18)(1942).

5. T. E. Beesley and R. P. W. Scott, Chiral Chromatography, John

Wiley and Sons, Chichester-New York, (1998)39.

6 R. P. W. Scott and T. E. Beesley, Analyst, 124(1999)713

7. R. P. W. Scott, Analyst, 124(1999).

8. E. Katz and R. P. W. Scott, J. Chromatogr., 253(1982)159.

9. R.P.W.Scott, Chromatography Detectors: Design. Function and

Operation, Marcel Dekker Inc., New York-Basle (1997).

10. G.J.Schmidt and R.P.W.Scott Analyst, 110(1985)757.

11. G.J.Schmidt and R.P.W.Scott, U.S. Patent 4,555,936 (1983)

12. K. K. Unger, Angew . Chem . Int . Engl . 11(1972)267.

13. M. Le Page, R. Beau, and J. Duchene, Fr. Pat.,

No 1.473,240 (1967).

14. M. Le Page and A. de Vries, Fr. Pat., No 1.475,929 (1967).

15. K. Unger and B. Scharf, Ger.. Pat., No 2,155,281 (1971).

16. D. W. Sindorf and G. E. Maciel, J. AM. Chem. Soc.

105(1983)1487.

17. D. W. Sindorf and G. E. Maciel, J. Phys. Chem. 87(1983)5516.

18. J.M.Vleeskens, Thesis, University of Technology , Delft,

The Netherlands,(1959).

19. J.H.de Boer and J.M.Vleeskens,Koninkl.ned.Akad.Wet.,

Proc.B61,2,(1957)85.

20. J.H.de Boer,J.H.Linsen and C.Okkerse,Koninkl.ned.Akad.Wet.,

Proc.B63,(1958)360.

21. J. Erkelens and B. G. Linsen J . Colloid Interface Sci .,

29(1969)464.

22. R. P. W. Scott and S. Traiman, J. Chromatogr .,196(1980)193.

23. K.K.Unger,"Porous Silica", J.Chromatogr. Libr.16(1979)99.

24. C. F. Simpson, U K Patent 1,310,8712,(1975).

25. C.H.Lochmüller and D.R.Wilder,J.Chromatogr. Sci.,



103

17(1979)574.

26. R.K.Gilpin and J.A.Squires, J.Chromatogr. Sci.,17(1979)574.

27. R.P.W.Scott and C.F.Simpson, J.Chromatogr. 197(1980)11.

28. R.P.W.Scott, J.Chromatogr., 142(1977)213.

29. R. Kunin, E. F. Meitzner and N. Bostnick, J. Am. Chem. Soc.

84(1962)84.

30. J. R. Millar, D. G. Smith, E. E. Marr and T. R. E. Kressman,

J. Chem. Soc.,218(1963).

31. R. Kunin, E. F. Meitzner, J. A. Oline, S. A. Fischer and N.W.

Frisch, Ind. Eng. Chem. Prod. Res. Develop.,1(1962)140.

32. K.K.Unger,"Porous Silica", Elsevier Scientific Publishing

Company, Amsterdam-Oxford-New York (1979)187.

33. S.G.Glasstone,"Textbook of Physical Chemistry, D.Van Nostrand,

New York,(1946)1198.

34. R. J. Laub and J. H. Purnell, J. Chromatogr., 112(1975)71.

35.E. D. Katz, K. Ogan and R. P. W. Scott, J.Chromatogr.

352(1986)67.

36. E. D. Katz, C. H. Lochmüller and R. P. W. Scott, Anal Chem.

61(1981)344.

37. W. H. Pirkle and D. W. House, J. Org. Chem., 44(1979)1957.

38. Y. Okamato, Chem. Lett., (1986)1237.

39. D. W. Armstrong, Y. Tang, S. Chen, Y. Zhou, C. Bagwill, and

J. R. Chen, Anal. Chem., 66(9)(1994)1473.

40. K. H. Ekborg-Ott, Chirality, (1998).

41. J. Haginaka and J. Wakai, Anal. Chem., 62 (1990)997.

42. A. M. Stalcup and K. L. Williams, J. Liq. Chromatogr.,

15(1)(1992)29.

R.P.W.Scott Liquid chromatography

Documents

Transcript of R.P.W.Scott Liquid chromatography